AU2016279798A1 - Glycopyrronium fatty acid salts and methods of making same - Google Patents

Abstract

Novel glycopyrronium fatty acid salts have been developed. Bi-phasic reaction conditions enable the desired counterion exchange reactions between glycopyrronium bromide and fatty acid salts of alkali metals and alkaline earth metals in methods to form glycopyrronium fatty acid salts. In preferred embodiments, an excess of the free fatty acid in the reaction mixture stabilizes the glycopyrronium fatty acid salt and reduces the formation of the impurity, Acid A. In some preferred embodiments, between 0.2 and 1.2 molar equivalent of excess free fatty acid is added to the reaction mixture. In another embodiment, approximately 1.2 molar equivalent of excess free fatty acid is added to the reaction mixture.

Description

inventions which were disclosed in Provisional Application Number 62/175,737, filed June 15, 2015, entitled

GLYCOPYRRONIUM FATTY ACID SALTS AND METHODS OF MAKING SAME. The benefit under 35 USC § 119(e) of the United States provisional application is hereby claimed, and the aforementioned application is hereby incorporated herein by reference.

FIELD OF THE INVENTION

The invention pertains to the field of glycopyrronium salts. More specifically, the 10 invention pertains to novel lipophilic glycopyrronium fatty acid salts.

BACKGROUND OF THE INVENTION

Glycopyrrolate (also known as glycopyrronium bromide) is a bromide salt with a quaternary ammonium counterion with the chemical name of 3[cyclopentyl(hydroxy)phenylacetoxy]-1,1-dimethyl pyrrolidinium bromide, a molecular 15 formula of Ci9H2sBrNO3 and a molecular weight of 398.34. Its chemical structure is shown in Table 2 below.

Trospium chloride is a quaternary ammonium salt with the chemical name of 3 (2 hydroxy-2,2 diphenylacetoxy)spiro[bicyclo[3.2.1]octane-8,l' pyrrolidin]-l'-ium chloride. The molecular formula of trospium chloride is C25H30CINO3 and its molecular weight is

427.97. The chemical structure of trospium chloride is:

WO 2016/204998

PCT/US2016/035967

The quaternary ammonium antimuscarinic drugs (QAAM) are particularly useful because of their ability to antagonize endogenous acetylcholine during periods of excessive acetylcholine production, or prolonged acetylcholine effect from physiologic and pharmacologic reasons. These compounds share the property that they do not appreciably penetrate the central nervous system (CNS), and glycopyrrolate and trospium chloride have been particularly useful in treating patients in need of a peripheral anticholinergic effect on antimuscarinic receptors.

The same biochemical property that is advantageous in preventing CNS distribution, also limits intestinal absorption, requiring the currently available formulations of these medications to be taken in the absence of food, and resulting in incomplete and variable bioavailability in patients.

SUMMARY OF THE INVENTION

Bi-phasic reaction conditions enable the desired counterion exchange reactions between glycopyrronium bromide and alkali and alkaline earth metal salts of fatty acids. Favorable partitioning of the glycopyrronium moiety into the organic phase (along with the fatty acids) and partitioning of the bromide into the aqueous phase preferably uses water and methyl tetrahydrofuran. While the glycopyrronium fatty acid salts are unstable with respect to hydrolysis under the reaction conditions and are isolated as oily products, an excess of the fatty acid in the reaction mixture stabilizes the glycopyrronium fatty acid salt and reduces the formation of the hydrolysis byproduct impurity, Acid A.

Methods of manufacturing glycopyrronium fatty acid salts preferably include use of a molar excess of the fatty acid. In some embodiments, at least a 0.2 molar equivalent of excess free fatty acid is added to the reaction mixture to form a glycopyrronium fatty acid salt. In some preferred embodiments, between 0.2 and 1.2 molar equivalent of excess free fatty acid is added to the reaction mixture. In another preferred embodiment, at least 0.6 molar equivalent of excess free fatty acid is added to the reaction mixture, in yet another preferred embodiment, between 0.6 and 1.2 molar equivalent of excess free fatty acid is added to the mixture to form a glycopyrronium fatty acid salt. In another

WO 2016/204998

PCT/US2016/035967 embodiment, approximately 1.2 molar equivalent of excess free fatty acid is added to the reaction mixture. In another embodiment, at least 1.1 molar equivalent of excess free fattyacid is added to the reaction mixture.

Since excess free fatty acids stabilize the formulations described herein, this may 5 result in enhanced bioavailability of the glycopyrronium moiety.

BRIEF DESCRIPTION OF THE DRAWINGS

Fig. 1 shows the HPLC Calibration Curve for glycopyrronium bromide by Bromide Peak.

Fig. 2 shows the HPLC Calibration Curve for glycopyrronium bromide by “Glycopyrronium” Peak.

Fig. 3 shows the HPLC Results of an exchange reaction of glycopyrronium bromide with potassium stearate.

Fig. 4 shows the HPLC Results of an exchange reaction of glycopyrronium bromide with potassium palmitate.

Fig. 5 shows the methylation of glycopyrrolate base with dimethyl carbonate.

Fig. 6 shows the HPLC calibration curve for the glycopyrrolate base.

Fig. 7 shows HPLC data for glycopyrronium stearate EE-008-008.

Fig. 8 shows HPLC data for glycopyrronium stearate EE-008-001-3B.

Fig. 9 shows gas chromatography data for glycopyrronium stearate sample EE-008-008.

Fig. 10 shows gas chromatography data for glycopyrronium stearate sample EE-008-00120 3B.

Fig. 11 shows NMR data for the whole spectrum of the glycopyrronium stearate sample EE-008-008.

WO 2016/204998

PCT/US2016/035967

Fig. 12A shows NMR data for a portion of the spectrum of Fig. 11.

Fig. 12B shows NMR data for a portion of the spectrum of Fig. 11.

Fig. 12C shows NMR data for a portion of the spectrum of Fig. 11.

Fig. 13 A shows NMR data for the whole spectrum of the Acid A by-product.

Fig. 13B shows NMR data for a portion of the spectrum of Fig. 13A.

Fig. 13C shows NMR data for a portion of the spectrum of Fig. 13A.

Fig. 14A shows NMR data for the glycopyrronium hydrolysis by-product, quaternary amino alcohol (QAA), residual water in DMSO-d6, s at 4.7 ppm.

Fig. 14B shows NMR data for a portion of the spectrum of 14A.

Fig. 14C shows carbon NMR data for the quaternary amino alcohol.

Fig. 14D shows carbon NMR data for a portion of the spectrum of Fig. 14C.

Fig. 15 A shows HPFC data for glycopyrronium laurate.

Fig. 15B shows an expanded view of the data of Fig. 15A, along with the area percent report.

Fig. 15C shows HPFC data for another run of the glycopyrronium laurate sample.

Fig. 15D shows an expanded view of the data of Fig. 15C, along with the area percent report.

Fig. 16A shows HPFC data for glycopyrronium palmitate.

Fig. 16B shows an expanded view of the data of Fig. 16A, along with the area percent report.

Fig. 16C shows HPFC data for another run of the glycopyrronium palmitate sample.

WO 2016/204998

PCT/US2016/035967

Fig. 16D shows an expanded view of the data of Fig. 16C, along with the area percent report.

Fig. 17A shows HPFC data for glycopyrronium linoleate.

Fig. 17B shows an expanded view of the data of Fig. 17A, along with the area percent report.

Fig. 17C shows HPFC data for another run of the glycopyrronium linoleate sample.

Fig. 17D shows an expanded view of the data of Fig. 17C, along with the area percent report.

Fig. 18 shows NMR data for glycopyrronium bromide, residual water in DMSO-d6 at 3.33 ppm, singlet.

Fig. 19 shows NMR data for glycopyrronium laurate.

Fig. 20 shows NMR data for glycopyrronium palmitate.

Fig. 21 shows NMR data for glycopyrronium linoleate.

Fig. 22 shows the glycopyrronium concentration versus glycopyrronium peak area.

Fig. 23 shows a blank chromatogram.

Fig. 24 shows a resolution solution chromatogram.

Fig. 25 shows ion chromatography data for a bromide standard.

Fig. 26 shows ion chromatography data for bromide in glycopyrronium stearate.

Fig. 27 shows ion chromatography data for potassium in glycopyrronium stearate.

Fig. 28 shows a standard gas chromatography chromatogram for stearic acid.

WO 2016/204998

PCT/US2016/035967

Fig. 29 shows a sample gas chromatography chromatogram for stearic acid analysis in glycopyrrolate stearate.

DETAILED DESCRIPTION OF THE INVENTION

Currently available quaternary ammonium anti-cholinergic muscarinic receptor antagonists compositions occur as a salt, with the quaternary ammonium cation and a nonorganic anion.

U.S. Patent No. 8,097,633, entitled Uses for Quaternary Ammonium Anticholinergic Muscarinic Receptor Antagonists in Patients Being Treated for Cognitive Impairment or Acute Delirium, issued January 17, 2012, and U.S. Patent Publication 2012/0088785, entitled New Uses for Quaternary Ammonium Anticholinergic Muscarinic Receptor Antagonists in Patients Being Treated for Cognitive Impairment or Acute Delirium, published April 12, 2012, both herein incorporated by reference, disclose methods for treating the adverse effects of acetylcholinesterase inhibitors using quaternary ammonium anti-cholinergic muscarinic receptor antagonists such as glycopyrrolate or trospium.

U. S. Patent 8,969,402, entitled Combined Acetylcholinesterase Inhibitor and Quaternary Ammonium Antimuscarinic Therapy to Alter Progression of Cognitive Diseases, issued March 3, 2015, and US Patent Publication No. 2013/0172398, entitled Combined Acetylcholinesterase Inhibitor and Quaternary Ammonium Antimuscarinic Therapy to Alter Progression of Cognitive Diseases, published July 4, 2013, both herein incorporated by reference, disclose administering quaternary ammonium anti-cholinergic muscarinic receptor antagonists in combination with acetyl-cholinesterase inhibitors to treat either cognitive impairment or acute delirium. This therapy results in a modification of a cognitive disorder or disease, namely a slow-down in the disease progression. New formulations for quaternary ammonium anti-cholinergic muscarinic receptor antagonists are also disclosed.

In preferred embodiments of the present invention, a quaternary ammonium anticholinergic muscarinic receptor antagonist includes a salt comprising an organic lipophilic

WO 2016/204998

PCT/US2016/035967 anion as an anionic component of the salt. In some preferred embodiments, the lipophilic anion of the quaternary ammonium anti-cholinergic muscarinic receptor antagonist preferably includes a fatty acid including at least eight carbon molecules. In some preferred embodiments, the quaternary ammonium anti-cholinergic muscarinic receptor antagonist is glycopyrrolate or trospium.

The quaternary ammonium antimuscarinic drugs (QAAM) are particularly useful because of their ability to antagonize endogenous acetylcholine during periods of excessive acetylcholine production, or prolonged acetylcholine effect from physiologic and pharmacologic reasons. These compounds share the property that they do not appreciably penetrate the central nervous system (CNS), and glycopyrronium bromide and trospium chloride have been particularly useful in treating patients in need of a peripheral anticholinergic effect on antimuscarinic receptors.

The same biochemical property that is advantageous in preventing CNS distribution, also limits intestinal absorption, requiring the currently available formulations of these medications to be taken in the absence of food, and resulting in incomplete and variable bioavailability in patients.

Enhancing the oral bioavailability of glycopyrrolate, trospium, and other QAAMs allows for administration of the medication without regard to food, and possibly without regard to other medications. It would also decrease the inter-subject variability in effect of the medication and reduce the effect of changes in gastrointestinal motility on drug absorption, because the degree of variability between patients is directly proportional to the time it takes to absorb a medication. There may also be an improvement in patient adherence to taking the medication by being able to take it in proximity to food.

A QAAM is produced with a lipophilic anion as the anionic component of a salt. Structure activity analysis (SAR) suggests that an optimum lipophilic anion would be a fatty acid of at least 8 carbon molecules, so that the hydrophobic tail of the molecules would provide enhanced lipid solubility to counteract the positive charge of the ionized cationic QAAM molecule. In some preferred embodiments, the appropriate salts come

WO 2016/204998

PCT/US2016/035967 from a family of medium and long chain fatty acids including, but not limited to: arachidic acid, stearic acid, palmitic acid, oleic acid, erucic acid, linoleic acid, arachidonic acid, lauric acid, capric acid, linolenic acid, or myristic acid.

The salt of the QAAM (cation) and the fatty acid (anion) can be produced through organic chemistry reactions referred to as ion swapping, counterion exchange or “salt metathesis. In such a reaction, the QAAM compound as the current elemental salt (glycopyrrolate hydrobromide, trospium chloride) is placed in a biphasic solution with the elemental salt of an omega-3 fatty acid such as α-linolenic acid. The solution is subjected to variations in temperature, pH and agitation to produce a salt that is selectively extracted into the organic phase. The extracts are concentrated under reduced pressure to remove the solvent and the salt is isolated, qualitatively and quantitatively identified and then stoichiometrically administered to animals. Quantitative serum and/or urine assays are used to make comparisons with the elemental salt of the QAAM, both in the presence and absence of food. Intravenous administration of the QAAM (example: glycopyrrolate) with quantitative serum and/or urine assay can be used as a reference standard for 100% bioavailability, and both the native compound and the synthesized salt are compared against this to establish their relative bioavailability in the presence and absence of food and other commonly co-administered medications.

In addition to the bioavailability studies mentioned above, quality studies would need to be done during the process to make sure there is no hydrolysis of the QAAM molecule in the process.

The synthesized fatty acids/ QAAM salt are useful as an individual product for the treatment of various diseases involving excessive acetylcholine activity in humans and animals, whether these are produced by a pathologic process or the use of a medication (including but not limited to: overactive bladder, sialorrhea, diarrhea, bradycardia, hyperhidrosis, overactive gastric secretion, dumping syndrome, bronchospasm, vasomotor rhinitis). Enhanced bioavailablity QAAM should be able to improve symptoms without causing significant central nervous system anti-cholinergic toxicity. Enhanced bioavailability allows for administration without regard to food, and may also allow for

WO 2016/204998

PCT/US2016/035967 transdermal absorption to be enhanced to the level that transdermal formulations of this product would be practical. This could also be used in combination with acetylcholinesterase inhibiting drugs, or any other drug that increases acetylcholine tone.

Bi-phasic reaction conditions enable the desired exchange reactions between glycopyrronium bromide and fatty acid salts of alkali metals and alkaline earth metals to produce glycopyrronium fatty acid salts. Favorable partitioning of the glycopyrronium moiety into the organic phase (along with the fatty acids) and partitioning of the bromide into the aqueous phase uses water and methyl tetrahydrofuran. While the glycopyrronium fatty acid salts are moderately unstable with respect to hydrolysis under the reaction conditions and are isolated as oily products, an excess of the fatty acid in the reaction mixture stabilizes the glycopyrronium fatty acid salt and reduces the formation of the hydrolysis byproducts, Acid A and the quaternary ammonium degradant (QAA).

The mixture of glycopyrronium fatty acid salt and excess free fatty acid may be isolated from each other to produce a consistent, well defined product. The glycopyrronium fatty acid salts described herein potentially offer the desired increase in glycopyrronium bioavailability. Since excess free fatty acids stabilize the formulations described herein, this may result in enhanced bioavailability of the glycopyrronium moiety.

Practical and scalable synthetic processes for preparation of lipophilic glycopyrronium fatty acid salts were developed. Initially, the preparation of the lipophilic glycopyrronium fatty acid salts from glycopyrronium bromide and fatty acid potassium salts was performed in organic solvents (anhydrous conditions) or biphasic (water/organic) conditions since this provided an opportunity for quick success in identifying a suitable preparative procedure. Solubility of the starting materials and products impacted these options, ultimately limiting the approach to bi-phasic (solvent/water) systems. The preparation of the lipophilic glycopyrronium fatty acid salts primarily utilized lauric acid, palmitic acid, linoleic acid and stearic acid (potassium stearate). The processes utilized different salts, including Na, K and Ca salts of fatty acids. In other embodiments, Mg or Ba salts of fatty acids are used.

WO 2016/204998

PCT/US2016/035967

The evaluation included multiple steps. The first step was screening and evaluation of lipophilic salt preparation options utilizing lauric acid. Subsequent process development and analytical method development were conducted using preparations using stearic acid. Then, at least 5 g of a lipophilic glycopyrronium fatty acid salt utilizing stearic acid was created. Three additional and different samples of at least 3 g lipophilic glycopyrronium salts were synthesized with additional fatty acids (palmitic acid, lauric acid, linoleic acid and stearic acid- see Table 23B below). The resulting four samples were then characterized.

Analytical methods were also developed for the glycopyrronium fatty acid salts. These methods appropriately and accurately assess the quality of future glycopyrronium fatty acid salts development samples. A more refined process for the general synthesis of glycopyrronium fatty acid salts has also been developed.

Some of the analytical method applications include, but are not limited to, an assessment of the salt exchange effectiveness, to ensure that the reactions are pushed to >95% conversion by determining an optimal number and amount of aqueous washes to remove inorganic bromide salts (byproducts), the determination of the optimal amount of excess free fatty acid (evaluating a suitable range of molar equivalents) needed to stabilize the active pharmaceutical ingredient (API) and the improvement of the isolation and purification of the reaction product (for example through precipitation with suitable solvent/antisolvent combinations) by evaluating the purity of the reaction products. In some preferred embodiments, the optimal number of aqueous washes is 3-4 washes. In other embodiments, the preferred number of washes is at least three washes.

Analytical method development also preferably includes methods for quantitation of chromophoric starting materials, products and degradants, method for quantitation of weakly- or non-chromophoric starting materials and degradants (fatty acids and derived salts, dimethylhydroxypyrrolidinium degradants), methods for quantitation of the bromide ion in the starting material and the product, and methods for quantitation of the potassium (or sodium) ion in the product. Individually, none of these methods alone would be sufficient to assess the effectiveness of the salt exchange process and the purity of the

WO 2016/204998

PCT/US2016/035967 derived products, but taken together, the combination of these orthogonal analytical methods achieves both of these objectives.

Some abbreviations used throughout this application are shown in Table 1 and chemical structures are shown in Table 2.

Table 1

API	Active Pharmaceutical Ingredient
ACN	Acetonitrile
Br	Bromine
DCM	Dichloromethane
DMSO	Dimethylsulfoxide
FA	Fatty acid
GC	Gas Chromatography
GP	Glycopyrronium
GPBr	Glycopyrronium Bromide
GPFA	Glycopyrronium Fatty Acid
HPLC	High Performance Liquid Chromatography
IPAc	Isopropyl acetate
K	Potassium
2-Me-THF	2-Methyl Tetrahydrofuran (Methyl-THF)
MIBK	Methyl Isobutyl Ketone
MTBE	Methyl-t-Butyl Ether
Na	Sodium
QAA	Quaternary Amino Alcohol (descriptive reference for degradant CAS 51052-74-5, 3-hydroxy-1,1dimethyl pyrrolidinium bromide)

Table 2

Entry	Chemical Structure	Origin	Notes
1	\—oh Z====s/V'°'v^\©x< \ Π I N. Br Vz Glycopyronnium Bromide mw 398.34	Starting Material
2	0 ® θΑ. MO R (M=K, Na)	Starting Material	Alkali metal salts of Fatty Acids

WO 2016/204998

PCT/US2016/035967

3	\—oh 0Ϊ° TX Λ	Reaction Product	Glycopyrronium Fatty Acid Salt(s)
4	Λ	Additive	Fatty Acids (excess)
5	Φ 0 M Br (e.g. KBr, NaBr)	Reaction By-Product	Bromide Salts
6	\—( OH ’’Acid A” CAS 427-49-6 a-cyclopentyl-a-hydroxy-benzene acetic acid	Degradant	Derived from Starting Material (entry 1) or Product (entry 3)
7	H°YX Λ Fatty acid salt of CAS 109357-53-1 3-hydroxy-1,1 -dimethyl pyrrolidinium Fatty acid salt	Degradant	Derived from Product (entry 3)
8	H°VA®/ θ I Nf_ Br =---/ “QAA”, (Quaternary Amino Alcohol), CAS 51052-74-5, 3hydroxy-1,1 -dimethyl pyrrolidinium bromide	Degradant	Derived from Starting Material (entry 1)
9	Methyl THF	Reaction Solvent

A general chemical structure for the targeted lipophilic glycopyrronium fatty acid salts 1 is shown below in Table 3A.

Table 3A

¹

Chemical structures for some preferred examples of glycopyrronium fatty acid salts are listed here.

R = C11H23 (lauric acid)

WO 2016/204998

PCT/US2016/035967

Glycopyrronium Laurate

R = C17H35 (stearic acid)

Glycopyrronium Stearate

R = C17H33 (oleic acid)

Glycopyrronium Oleate

Glycopyrronium Palmitate

Glycopyrronium Caprate

R = C19H31 (arachidonic acid)

WO 2016/204998

PCT/US2016/035967

Glycopyrronium Arachidate

Glycopyrronium Linoleate

Glycopyrronium Myristate

R = C17H29 (α-linolenic acid)

WO 2016/204998

PCT/US2016/035967 /''v

Glycopyrronium a-Linolenate

R = C17H29 (γ-linolenic acid) /X

Glycopyrronium y-Linolenate

There were two preferred starting materials (glycopyrronium bromide 2 and glycopyrrolate base 3), as shown in Table 3B below.

Table 3B

3

Synthesis of Glycopyrronium salts

A practical, scalable synthesis of glycopyrronium salts of fatty acids was developed. The general chemical structure of the glycopyrronium salt of fatty acids is shown as compound 1 in Table 3A. Three different approaches were considered, starting from either glycopyrronium bromide (compound 2 in Table 3B, also known as “glycopyrrolate”) or from a synthetic precursor, designated herein as the glycopyrrolate base (compound 3 in Table 3B). The three approaches were a) salt metathesis using ion exchange resins, starting from glycopyrronium bromide (compound 2 in Table 3B), b) direct salt metathesis, starting from glycopyrronium bromide (compound 2 in Table 3B),

WO 2016/204998

PCT/US2016/035967 and c) synthesis via glycopyrronium methyl carbonates, using glycopyrrolate base (compound 3 in Table 3B) as the starting material.

Adding at least 0.2 molar equivalents of excess free fatty acids into a preparation with fatty acid salts and glycopyrronium bromide in a water/Me-THF system stabilized the reaction mixture and allowed for the formation of glycopyrronium fatty acid salts.

Free fatty acid as defined herein is fatty acid in its free form, which is different from the fatty acid in its ionized form (salt form).

in preferred embodiments, at least a 0.2 molar equivalent of excess free fatty acid is added to the reaction mixture to form a glycopyrronium fatty acid salt. In some preferred embodiments, between 0.2 and 1.2 molar equivalent of excess free fatty acid is added to the reaction mixture. In another preferred embodiment, at least 0.6 molar equivalent of excess free fatty acid is added to the reaction mixture. In vet another preferred embodiment, between 0.6 and 1.2 molar equivalent of excess free fatty acid is added to the mixture to form a glycopyrronium fatty acid salt. In another embodiment, approximately 1.2 molar equivalent of excess free fatty acid is added to the reaction mixture. In another embodiment, at least 1.1 molar equivalent of excess free fatty acid is added to the reaction mixture.

in preferred embodiments, the isolated glycopyrronium fatty acid salt mixtures have an enrichment of the fatty acid (relative to glycopyrronium) compared to the input ratio, in embodiments using 1.2 molar equivalent of excess free fatty acid, the isolated products preferably have more than the 2.2:1 FA:GP input ratio. In some of these embodiments, the ratio is approximately between 2.25:1 and 3.00: i.In some of these embodiments, the ratio is approximately between 2.29:1 and 2.87:1.

In one example, a reaction mixture of glycopyrronium bromide with potassium laurate is stabilized with respect to the formation of the by-product (Acid A) by adding excess free lauric acid. Some preferred fatty acids include, but are not limited to, arachidic acid, stearic acid, palmitic acid, oleic acid, erucic acid, linoleic acid, arachidonic acid, lauric acid, capric acid, linolenic acid, or myristic acid. Some preferred salts for the fatty

WO 2016/204998

PCT/US2016/035967 acid salt include, but are not limited to, Na, K, or Ca salts. In other embodiments, Mg or

Ba salts may be used.

As described herein, excess free fatty acid is relative to the glycopyrronium bromide and fatty acid salt used. For example, for a lauric acid reaction with potassium, it is the excess free lauric acid relative to the glycopyrronium bromide and potassium laurate used.

The excess free fatty acid stabilizes the reaction mixture. The larger excesses of free fatty acid (0.6-1.2 molar excess) improved the phase separations, improved stability of the organic extract solutions, and improved stability of the isolated products.

There is potential to isolate a mixture of glycopyrronium fatty acid salt and excess free fatty acid to ensure a consistent, well defined product.

Ion exchange

From the outset, the synthesis approach using ion exchange resins was challenging due to the known hydrolytic instability of glycopyrronium bromide at pH values above pH 5.6 (see, for example G Gupta, V.D., Stability of Oral Liquid Dosage Forms of Glycopyrrolate Prepared With the Use of Powder, International Journal of

Pharmaceutical Compounding, 2003, 7(5), 386-388, herein incorporated by reference) and the complexity and cost of the required technology. While aqueous solutions of glycopyrronium bromide are reasonably stable at ambient temperature at pH 5.6 and below, glycopyrronium bromide would be expected to rapidly hydrolyze at the ester linkage at the pH values required in a salt exchange process employing resins.

Starting from commercially available glycopyrronium bromide (structure 2 in Table 3B), the corresponding fatty acid salts (structure 1 in Table 3A) can in principle be derived through the use of anionic ion exchange resins. Table 4 provides an ion exchange flow diagram, which shows a schematic of how this may be accomplished.

Table 4

WO 2016/204998

PCT/US2016/035967

However, the ion exchange approach using resins requires expertise beyond standard synthesis methodologies. Besides the selection of a suitable resin, a number of parameters need to be carefully chosen to maximize the effectiveness of this process.

Factors that impact the adsorption-desorption isotherms in ion exchange include, but are not limited to, the relative ratio of substrate to resin, the solvent system (eluent), the temperature, and the substrate concentration. The hydrophobic interactions between the resin and the fatty acid backbone may play an important role as well in binding efficiency (besides the electrostatic attraction of the polar groups). For this reason, the resin choice itself may be an important consideration (see, for example Ihara, Y. Adsorption of Fatty Acid Sodium Salts on Ion Exchange Resins, Journal of Applied Polymer Science, 1986, 32(6). 5665-5667, herein incorporated by reference). In Ihara, the weakly basic resin with hydrophobic phenyl groups in the polymer (IRA94) showed a better performance than the less hydrophobic resin (IRA68). The adsorption efficiency of the fatty acid sodium salts dramatically increased, in going from C-6 to C-12 fatty acid salts.

An additional challenge with the ion exchange methodology is the instability of glycopyrrolate at elevated pH in aqueous solutions due to ester hydrolysis. This compound

WO 2016/204998

PCT/US2016/035967 is reasonably stable at ambient temperature at pH 5.6 and below (Gupta, 2003) but is expected to be susceptible to hydrolysis of the ester under the processing conditions (as pH values will be substantially higher than pH 5.6).

The other two approaches were evaluated for the preparation of lipophilic glycopyrronium fatty acid salts.

Laurie acid was selected as a model fatty acid for the initial work. The exchange reaction of glycopyrronium bromide with potassium laurate was screened in various solvent systems.

The best solvent system was water / methyl tetrahydrofuran (Me-THF) which provided the desired product, glycopyrronium laurate. Unfortunately, the isolated oily product was unstable and decomposed into the by-product, Acid A. The decomposition products from attempted methylation of the glycopyrrolate base with dimethyl carbonate, which include Acid A and the methyl ester of Acid A, are shown in Table 5.

Table 5

Potassium stearate and potassium palmitate were also investigated in the MeTHF/water system in attempts to improve the stability of the target salts and to obtain the target materials in a solid form. These reactions failed to provide solid product(s) and showed a partial hydrolysis to Acid A during the process.

Direct Salt Metathesis (counterion exchange)

The reaction of glycopyrronium bromide with fatty acid salts was performed in organic solvents (anhydrous conditions) and biphasic (water/organic) conditions. The reactions in organic solvents did not result in any product formation, while using biphasic conditions provided the desired products in the reaction mixture.

WO 2016/204998

PCT/US2016/035967

The general chemistry of this process is shown below:

Glycopyrronium Fatty Acid Salt

An example of the chemistry in this process is shown below:

Example

/

KOCOCuHja

Solvents

Br + KBr h

OCOCt 1 H23

A direct salt metathesis (“ion swap”) between glycopyrronium bromide and the sodium or potassium salt of representative fatty acids using an extractive process was successfully developed. In this process, the water soluble inorganic salt derived from the bromide exchange with the carboxylate conjugate base of the fatty acid is partitioned in the aqueous phase and removed from the reaction medium to drive the exchange process.

A direct counterion exchange between glycopyrronium bromide and suitable fatty acid salts was then evaluated. The driving force for a typical salt metathesis reaction is the generation of an insoluble salt that precipitates out of solution. In the case of fatty acid salts, the most likely starting materials would be the corresponding silver salts. For example, see US patent publication 2011/0306650, published December 15, 2011, and incorporated herein by reference, which describes the preparation of glycopyrronium chloride by reacting glycopyrronium bromide with silver acetate to produce glycopyrronium acetate (silver bromide precipitates out). The acetate salt is then reacted with HC1 to produce the glycopyrronium chloride (along with acetic acid).

Glycopyrronium sulfate is produced similarly by salt metathesis using silver sulfate. Such starting materials were not desirable due to concerns over heavy metal residues in the drug substance, as they were expected to be difficult to fully remove from the desired

WO 2016/204998

PCT/US2016/035967 glycopyrronium fatty acid salts and require special waste disposal considerations for the precipitated silver salts on large scale (preferably via metal recovery).

Nonetheless, a proper solvent choice in combination with the fatty acid salts of innocuous metals such as sodium, potassium or calcium was considered for assessment of counterion exchange through precipitation of the derived bromide salts (sodium, potassium or calcium bromide). Alternatively, the counterion exchange reaction could be driven by extracting the highly soluble inorganic salt into the aqueous phase.

In order to select the reaction solvents, the glycopyrronium bromide (GPBrstructure 2 in Table 3B) solubility in various solvents was determined using HPLC calibration curves (see Figs. 1 and 2). The results of glycopyrronium bromide solubility in various solvents at room temperature are summarized in Table 6 below. The experiment was conducted with 50 mg of glycopyrronium bromide and 1 mL of solvent in most cases. The mixture was stirred at room temperature for 24 hours and then filtered to afford a clear filtrate. The filtrate was diluted with acetonitrile and checked by HPLC.

Table 6

Solvents	At Room Temperature	Note
Water	>840 mg/mL	Additional GPBr was used
Methanol	>541 mg/mL	Additional GPBr was used
Ethanol	142 mg/mL	Additional GPBr was used
IPA	26 mg/mL
2-Butanol	24 mg/mL
ieri-Butanol	4.8 mg/mL
Acetone	5.4 mg/mL
MEK	0.4 mg/mL
MIBK	<0.1 mg/mL	Out of calibration limit
THF	<0.1 mg/mL	Out of calibration limit
Me-THF	<0.1 mg/mL	Out of calibration limit
DCM	31 mg/mL

Three lauric acid salts (Na, K, and Ca) were prepared as model salts. These salts were prepared following the procedure in Zacharie et al. (A Simple and Efficient Large21

WO 2016/204998

PCT/US2016/035967

Scale Synthesis of Metal Salts of Medium-Chain Fatty Acids, Organic Process Research & Development 2009, 13, 581-583), herein incorporated herein by reference.

Preparation of Sodium Laurate

Laurie acid (50.08g, 1.0 eq.) was dissolved in ethanol (500mL, 10vol., 95% denatured) at room temperature and reacted with NaHCO₃ (18.9g, 0.9 eq.) in ethanol (500mL, 10vol., 95% denatured) at refluxing temperature. The reaction mixture (suspension) was stirred at refluxing temperature (~77°C) overnight. Some solids precipitated overnight (vs. solution reported in literature) and the mixture was diluted with additional ethanol (-1.5L). The solution was decanted and cooled to room temperature over four hours (solid product precipitated at ~55°C). The slurry was filtered and the product was washed with MTBE (3x200mL) to remove excess free un-reacted lauric acid. The white wet cake was dried in the air over the weekend and in a vacuum oven at 50+5°C overnight to afford 27.2g (54.4% yield) of white solid. While the yield was lower than in the prior art literature, the procedure could be optimized for better yield. Some methods for improving yield include, but are not limited to, optimizing the concentration during the crystallization process or using an antisolvent to decrease the solubility of the product and improve the yield.

Preparation of Potassium Laurate

The second experiment was conducted with 25g (1.0 eq.) of lauric acid and 11.25g of KHC0₃ (0.9 eq.) in ethanol (250 mL) at reflux. The product was crystallized from ethanol/MTBE (1/1) to afford 22.1 g (82.5% yield) of potassium laurate after drying in the air over the weekend and in a vacuum oven at 50+5°C overnight.

In order to select the reaction solvents and their optimized volumes, the solubility of potassium laurate in various solvents was checked by adding solvent to dissolve the solids under sonication. A vial was charged with potassium laurate (~100mg) and then solvent was added dropwise under sonication. The potassium laurate could not be dissolved with most solvents (200 vol.). The results of potassium laurate solubility in various solvents are summarized in Table 7 below:

WO 2016/204998

PCT/US2016/035967

Table 7

Solvents	At Room Temperature	At 50°C
Water	-138 mg/mL	NA
Methanol	-50 mg/mL	NA
Ethanol	-14 mg/mL	NA
IPA	< 5mg/mL	< 5 mg/mL
Methyl Acetate	< 5mg/mL	< 5 mg/mL
Ethyl Acetate	< 5mg/mL	< 5 mg/mL
Isopropyl Acetate	< 5mg/mL	< 5 mg/mL
Butyl Acetate	< 5mg/mL	< 5 mg/mL
Acetone	< 5mg/mL	< 5 mg/mL
MEK	< 5mg/mL	< 5 mg/mL
MIBK	< 5mg/mL	< 5 mg/mL

Preparation of Calcium Laurate

The third experiment was conducted with 15.6g (1.0 eq.) of lauric acid and 2.6g of 5 Ca(OH)2 (0.9 eq.) in ethanol (450 mL) at reflux for 4 hours. The product precipitated at reflux temperature. The reaction mixture was further diluted with methanol (500mL) at reflux temperature. The slurry was cooled to room temperature and filtered then rinsed with an MTBE wash. The solid was dried in the air overnight and in a vacuum oven at 50+5°C overnight to afford 10.4g of white solid (67.7% yield). The procedure could be optimized to increase yield. Some methods for improving yield include, but are not limited to, optimizing the concentration during the crystallization process or using an antisolvent to decrease the solubility of the product and improve the yield.

Reaction in Organic solvents

Salt metathesis via precipitation was attempted using potassium laurate, sodium laurate, potassium stearate and potassium palmitate, in combination with acetone, methanol and dichloromethane as solvents.

Potassium laurate was selected as the model fatty acid salt for this approach. The first attempted reaction was conducted in anhydrous acetone. Glycopyrronium bromide (0.508g, 1.0 eq.) was charged into a 250mL flask followed by acetone (50mL). Potassium laurate (1.1 eq.) was added to the mixture. The mixture was stirred and additional acetone

WO 2016/204998

PCT/US2016/035967 (lOOmL) was added in an attempt to dissolve the solids. The solids failed to dissolve completely. The mixture was stirred overnight and was then concentrated to a residue. Ethyl acetate (EtOAc, lOOmL) was used to extract the crude residue. The extract was concentrated to afford only 48mg (less than 10% wt recovery) of an oily residue. HPLC analysis of the residue showed that several peaks were present. Proton NMR analysis showed that a complex mixture was obtained.

The reaction failed to provide the desired product, most likely due to the low solubility of both the glycopyrronium bromide and the potassium laurate.

Since both starting compounds are soluble in methanol, the next reaction was attempted in methanol. Glycopyrronium bromide (1.036g, 1.0 eq.) was charged into a 20mL vial followed by methanol (5mL) to dissolve all solids. A potassium laurate solution in methanol (0.691g, 1.1 eq., lOmL of methanol) was added to the mixture. The mixture was stirred over the weekend at ambient temperature and then the solution was checked by HPLC which showed that one new peak (RRT=1.37) was formed at -47% HPLC AUC. The solution was concentrated to remove methanol and then extracted with EtOAc (3 x 100mL). The EtOAc extracts were combined and concentrated to an oily residue (0.8g). HPLC analysis showed that the new peak was enriched to 61% AUC. Proton NMR analysis showed that a methyl ester peak at 3.77ppm was formed. The proposed structure for this by-product follows:

Me

Proposed by-product

The solid (0.87g) remaining after the EtOAc extraction was checked by HPLC which showed that the major peak was “Bromide” with-91% AUC. However, proton NMR analysis suggested that the solid was a mixture of at least three possible compounds listed in Table 8 below.

Table 8

WO 2016/204998

PCT/US2016/035967

'—Θ Br Proposed by-Product

ΗΟχ^/ '-'y®gCOC11H23 Proposed by-Product

K gCOCuHsg

The glycopyrronium fatty acid salt was unstable in the methanolic reaction mixture. However, it was confirmed that the starting material, glycopyrronium bromide (structure 2 in Table 3B), in the absence of other components, is stable in methanol.

To explore the exchange reactions under anhydrous conditions, four fatty acid salts (potassium laurate, sodium laurate, potassium stearate, and potassium palmitate) were tested by suspending the reactants in dichloromethane (DCM) and stirring for 90h at room temperature. Any precipitation of potassium bromide (or sodium bromide) was expected to provide enrichment of the glycopyrronium peak relative to the bromide peak in the

HPLC analysis of the filtered DCM solutions. The reaction mixture (DCM, 1.0 mL) was filtered through a syringe filter, concentrated, and the resulting solid residue was dissolved in a mobile phase and analyzed by HPLC. The HPLC results of the exchange reactions in an anhydrous DCM system are shown in Table 9 below.

Table 9

Sample Name	Sample Information	HPLC area ratio of “Bromide”/“Glycopyrronium”	Comments
GPBr	API (Bromide Salt)	1:18	Standard Solution
XL-007-102	Potassium Laurate	1:15.6	No loss of bromide
XL-007-103	Sodium Laurate	1:17.8	No loss of bromide
XL-007-104	Potassium Stearate	1:17.0	No loss of bromide
XL-007-105	Potassium Palmitate	1:16.4	No loss of bromide

WO 2016/204998

PCT/US2016/035967

The bromide /glycopyrronium ratio clearly indicated that no enrichment of glycopyrronium relative to bromide had taken place. No desired product was observed using anhydrous conditions (DCM) with potassium laurate, sodium laurate, potassium stearate or potassium palmitate.

All attempts were unsuccessful due to solubility issues (low solubility of glycopyrronium bromide in acetone), instability (transesterification reaction in methanol) and insufficient solubility difference (in dichloromethane).

Reaction in Biphasic conditions

An alternative approach whereby the exchange process is accomplished in a A counterion exchange via selective biphasic solvent mixture was more successful, extraction is shown in Table 10.

Br®

GP Bromide (Glycopyrrolate)

Solvents

R = lauric (C-12) palmitic (C-16 ) Stearic (C18)

M = Na, K

Best solvent: MeTHF

Six solvents were evaluated in the extractive procedure: 2-methyl tetrahydrofuran (MeTHF), methyl tert-butyl ether (MTBE), isopropyl acetate (IPAc), methyl isobutyl ketone (MIBK), toluene, and dichloromethane (DCM). Using sodium laurate as the model fatty acid salt, MeTHF was determined to be the best solvent, affording an organic phase that was highly enriched in glycopyrronium laurate.

Further experiments using potassium laurate, potassium palmitate and potassium stearate demonstrated the generality of this methodology. The reaction products were characterized by proton NMR analysis and confirmed to be consistent with the fatty acid

WO 2016/204998

PCT/US2016/035967 salts. HPLC analysis using a developmental HPLC method showed enrichments of up to

96% area/area following 3-4 aqueous washes of the MeTHF extracts.

A hydrolysis side reaction was observed to be a significant problem in the counterion exchange process at the prevailing pH (8.0 - 8.3). This issue was resolved by using an excess of the free fatty acid to “buffer” the medium and minimize the hydrolysis reaction, and by performing the exchange and further processing of the materials at ambient temperature.

Since both starting compounds are very soluble in water, the reaction was attempted in water/Me-THF. Glycopyrronium bromide (structure 2 in Table 3B) (1.0 eq.) was charged into a 20mL vial followed by water (2mL) to dissolve all solids. The glycopyrronium bromide solution was transferred into a solution of potassium laurate in water (1.1 eq., 5mL of water) in a 20mL vial. The glycopyrronium bromide vial was rinsed with water (3xlmL) and the rinse was transferred into the reaction vial. HPLC analysis showed that the area ratio of the “Bromide” peak to the “Glycopyrronium” peak was 20/80. Me-THF (lOmL) was added to the reaction solution. The mixture was stirred for one hour and then settled for phase separation. Both layers were checked by HPLC and it was found that the area ratio of the “Bromide” peak to the “Glycopyrronium” Peak was significantly different between the two layers (XL-007-071-1 and XL-007-071-2, Table 11 below). The reaction mixture was then re-mixed for one hour and then settled for one hour to afford two layers (XL-007-071-3 and XL-007-071-4). HPLC analysis showed that the ratio of the two peaks was unchanged. The results showed that more “Bromide” was contained in the aqueous layer and the “Glycopyrronium” was extracted into the Me-THF layer. The aqueous layer was removed, replaced with fresh water, mixed and then settled to afford two layers (XL-007-071-5 and XL-007-071-6). HPLC analysis showed that the Me-THF layer contained 93 area% of “Glycopyrronium”. The Me-THF layer was washed with water an additional two times and the product in the Me-THF layer was enriched to 96 area% (XL-007-071-10). The HPLC results for the aqueous and organic layers of the reaction mixture are shown in Table 11.

WO 2016/204998

PCT/US2016/035967

Table 11

Sample Name	Sample Information	HPLC Area Ratio of “Bromide”/”Glycopyrronium”
XL-007-071-0	Initial Water solution of both components	20/80
XL-007-071-1	Aqueous Layer	80/20
XL-007-071-2	Me-THL Layer	11/89
XL-007-071-3	Aqueous Layer	80/19	Repeated
XL-007-071-4	Me-THL Layer	11/89	Repeated
XL-007-071-5	Aqueous Layer	68/32	After first water wash
XL-007-071-6	Me-THL Layer	7/93
XL-007-071-7	Aqueous Layer	27/73	After second water wash
XL-007-071-8	Me-THL Layer	6/94
XL-007-071-9	Aqueous Layer	7/93	After third water wash
XL-007-071-10	Me-THF Layer	4/96*

One aliquot sample from the washed Me-THF layer (XL-007-071-10) was taken and concentrated to an oil which was analyzed by proton NMR in different solvents (CDCI3, DMSO-dg, and D₂O) for comparison against the starting materials. The proton NMR spectrum was consistent with the expected spectrum of the desired product:

OCOC-] -] H23

To improve upon the selective solubility and partitioning of bromide and fatty acid salt between the aqueous and solvent layers, the reaction of glycopyrronium bromide (structure 2 in Table 3B) with sodium laurate was screened in the following six solvents with water: Methyl tetrahydrofuran (Me-THF), Methyl tert-butyl ether (MTBE), Isopropyl acetate (IPAc), 4-Methyl-2-pentanone(MIBK), Toluene, and Dichloromethane (DCM).

The reaction was conducted with glycopyrronium bromide (lOOmg, 1.0 eq.) and sodium laurate (61 mg, 1.1 eq.) with 7 mL of water and 7 mL of organic solvent at room temperature. The reaction mixture was stirred over four days and then settled. Both layers

WO 2016/204998

PCT/US2016/035967 were checked by HPLC and the analytical results for the sodium laurate/glycopyrronium bromide exchange are summarized in Table 12 below:

Table 12

Experiment	Solvent System	HPLC Area Ratio of “Bromide”/” Glycopyrronium”
		Organic Layer	Aqueous Layer
1	Me-THF/water	18/82L1J	46/54L1J
2	MTBE/water	25/75L2J	30/70L1J
3	IPAc/water	80/20121	30/70L1J
4	Toluene/water	40/60121	26/74L1J
5	MIBK/water	72/28121	29/71L1J
6	DCM/water	65/35121	38/62L1J

[1] Decomposed product (Acid A) was observed in significant amounts by HPLC [2] HPLC peak areas were very small.

From these six experiments, Me-THF was the best solvent for selective product partition and extraction. Unfortunately, the decomposition product (Acid A, RRT=0.74) was observed in both the organic layer and aqueous layer.

Acid A

Me-THF was selected for further investigation based on the results of the solvent screening. The reaction using potassium laurate and Me-THF was scaled up to 4.89g of glycopyrronium bromide with potassium laurate (3.07 g, 1.05 eq.) in 25mL of water and 53mL of methyl THF at room temperature. The mixture was stirred for one hour and settled for phase separation. Three layers formed and each was checked by HPLC and the results of the scale up reaction in water/Me-THF are summarized in Table 13. It was confirmed that an additional peak (t_R =5.29 min, RRT=0.74) was detected and that this peak increased over time. This peak was confirmed to be the by-product, Acid A, by comparison with a separate preparation and isolation of Acid A. Acid A, the hydrolysis

WO 2016/204998

PCT/US2016/035967 product of glycopyrrolate base (structure 3 in Table 3B) was isolated and characterized by

HPLC and NMR (*H and ¹³C).

The bottom layer (aqueous) was removed and the top two layers were washed with water (3x20mL). It was observed that only two layers were formed after the first phase separation. The final organic layer (XL-007-080-4A) was concentrated at 60°C on a rotary evaporator to afford a residue which was re-dissolved with methyl THF (lOOmL). The insoluble solids (KBr and potassium laurate) were filtered and dissolved with water and then checked by HPLC (XL-007-080-P1). The filtrate was concentrated to dryness to afford an oil (6.12g, XL-007-080-P2) and HPLC analysis showed that the bromide peak was reduced to 1.34 area%. The product was dried in a vacuum oven at 50°C over the weekend. NMR analysis of the dried product showed some decomposition. The dried oily product was dissolved with MTBE (100 mL) and precipitation was attempted by adding n-heptane, resulting in two liquid layers. Both layers were checked by HPLC (Top layer: XL-007-082-1 and bottom layer: XL-007-082-2) and showed high levels of the by15 product, Acid A.

Table 13

Sample	HPLC Results (AUC)
ID	Information	Bromide	Acid A	Glycopyrronium
XL-007-080- 1A	Top Layer	13.57%	0.51%	85.92%
XL-007-080- 1B	Middle Layer	14.37%	0.50%	85.13%
XL-007-080- 1C	Bottom Layer	85.61%	ND	14.39%	Phase Removed
XL-007-080- 2C	Bottom Layer	76.66%	ND	23.34%	Phase Removed
XL-007-080- 3A	Top Layer	3.65%	1.44%	94.91%
XL-007-080- 3C	Bottom Layer	62.93%	ND	37.07%	Phase Removed
XL-007-080- 4A	Top Layer	1.85%	1.83%	96.32%
XL-007-080- 4C	Bottom Layer	24.96%	2.80%	71.23%	Phase Removed

WO 2016/204998

PCT/US2016/035967

XL-007-080- P1	Filtered Solids	12.88%	ND	ND	Potassium Laurate as major peak (lower response)
XL-007-080- P2	Oil Product	1.34%	4.84%	93.82%
XL-007-082-1	Top Layer	0.83%	12.37%	86.80%
XL-007-082-2	Bottom Layer	1.01%	15.06%	83.94%

It was confirmed that the isolated oily product had partially decomposed. Most likely, this decomposition is consistent with a hydrolysis reaction, which is facilitated at elevated temperature and elevated pH. The proposed decomposition reaction scheme is shown below:

(UV active) (UV inactive)

Glycopyrronium

Cation A Anion A h₂o

H0\--N®xCH3 UV ch3	vaph θ o	Ci 1H23CO2H
(UV inactive)	(UV active)	(UV inactive)

Cation B Anion B

Free Fatty Acid

The nature of the reaction mixture, which included components with strong 10 chromophores and others that had weak chromophores (for all intents and purposes herein labeled as “UV inactive”) rendered the accurate analysis and quantitation of all components in the mixtures very challenging. For this reason, orthogonal and complementary analytical methods were developed to facilitate both the in-process analysis of the reactions and the assessment of both the purity and the composition of the isolated reaction products.

The desired product, glycopyrronium laurate, was isolated by Me-THF extraction from an aqueous solution of glycopyrronium bromide and potassium laurate. Unfortunately, the isolated product was unstable during the isolation process with the levels of Acid A increasing over time. The isolated product degraded at 50°C from ~94

WO 2016/204998

PCT/US2016/035967 area% to 83-86 area% over a weekend. Glycopyrronium bromide itself was stable in the

HPLC diluent solution (ACN/water) at ambient temperature over two weeks (by HPLC).

The preparation of glycopyrronium laurate was repeated at 5.0g scale (GPBr). The HPLC results of the repeated scale-up reaction in water/ME-THF are summarized in Table

14 below. It was confirmed by HPLC that the decomposition product Acid A, was formed during the work up and isolation.

Table 14

Sample	HPLC Results (AUC)
ID	Information	Bromide	Acid A	Glycopyrronium	PH
XL-007-091-A	First aq. Layer	84.70%	0.78%	14.52%	8.00
XL-007-091-B	Second aq. Layer	74.82%	1.46%	23.72%	8.00
XL-007-091-C	Third aq. Layer	57.96%	2.97%	39.08%	8.07
XL-007-091-D	Fourth aq. Layer	19.51%	4.90%	75.59%	8.25
XL-007-091-1	Me-THF Layer (final)	4.31%	4.12%	91.06%
XL-007-091-2	Me-THF layer after overnight ambient	3.25%	6.07%	90.67%

The Me-THF solution was stress-tested at room temperature and at 50°C over the 10 weekend with and without the addition of an excess of free lauric acid. The results of the stress test are listed in Table 15. While the room temperature solutions showed far less degradation than the heated samples, decomposition was still significant. The attempt to stabilize the product with excess free lauric acid resulted in less decomposition being observed than in the corresponding samples without excess free lauric acid. It was noted that the Me-THF solutions split into two layers at 50°C. This was likely due to a decrease in the solubility limit of water in the mixture upon heating.

Without excess free lauric acid, the product slowly decomposed at ambient temperature from 90.67 area% product content to 87.72% in the Me-THF solution. At 50°C, the product decomposed significantly, to only 57 area% in the top layer and -50% in the bottom layer. In the presence of excess free lauric acid at ambient temperature, the

WO 2016/204998

PCT/US2016/035967 product showed only slight decomposition (Expt. 2A) while at 50°C, the product decomposed, leaving only -65-70 area%, even with excess free lauric acid present.

Table 15

Expt.	Conditions		HPFC Results (AUC%)
Bromide	Acid A	Glycopyrronium
0	Initial Me-THF Solution		3.25%	6.07%	90.67%
1A	Me-THF Solution at RT over weekend		3.02%	9.26%	87.72%
IB	Me-THF Solution at 50°C over weekend	Top	2.79%	40.08%	57.13%
Bottom	4.07%	46.08%	49.47%
2A	Me-THF Solution in the presence of Fauric Acid at RT over weekend		3.95%	6.61%	89.44%
2B	Me-THF Solution in the presence of Fauric Acid at 50°C over weekend	Top	2.07%	28.36%	69.58%
Bottom	4.38%	30.01%	65.60%

The desired product in Me-THF degraded over time. The product degraded significantly faster at 50°C. With the presence of excess free lauric acid (approximately 1.0 molar equivalent excess), the product decomposition was slower especially at ambient temperature.

A Me-THF extraction of the aqueous solution of glycopyrronium bromide and potassium laurate produced the desired product but the desired product was unstable during the extraction process at room temperature. The observed decomposition product was confirmed by independent preparation and isolation to be the hydrolysis product (Acid A). It was confirmed that the product decomposed significantly faster at 50°C. The isolated crude product was oily and failed to precipitate from heptane and other solvents (MTBE, IPAc, and EtOAc).

Since the oily product obtained from potassium laurate was unstable, fatty acids with longer chains were tested in an attempt to get solid products with potentially improved stability.

WO 2016/204998

PCT/US2016/035967

Potassium stearate and potassium palmitate were prepared and tested with glycopyrronium bromide in the Me-THF/water system.

Stearic acid and palmitic acid were used as the potassium salts. The preparations of potassium stearate and potassium palmitate were performed following the procedures from Zacharie et al. (A Simple and Efficient Large-Scale Synthesis of Metal Salts of Medium-Chain Fatty Acids, Organic Process Research & Development 2009, 13, 581— 583), which are summarized below.

For the preparation of potassium stearate, stearic acid (25.Og, 1.0 eq.) was dissolved in ethanol (500mL, 20vol., 95% denatured) at 45 °C in a 2-L round bottom flask. KHCO3 (7.88g, 0.9 eq.) was added into the solution and then the reaction mixture was heated to reflux (~77°C). The reaction was stirred overnight (21 hours) at reflux. MTBE (500mL) was added to the solution at 65 °C. Some foaming occurred and vigorous stirring was needed to break the foam. A second portion of MTBE (500mL) was then added at 50°C and the resulting slurry was cooled to room temperature over approximately 3 hours. The slurry was filtered and the wet cake was washed with MTBE (3xl25mL). The white wet cake was dried in a vacuum oven at ~50°C overnight to afford a white solid product (25.4g, quantitative yield).

For the preparation of potassium palmitate, palmitic acid (15.0g, 1.0 eq.) was dissolved in ethanol (200mL, 95% denatured) at 40°C in a 1-L round bottom flask. Solid KHCO3 (5.27g, 0.9 eq.) was added into the solution and then the reaction mixture was heated to reflux (~77°C). The reaction was stirred overnight (20 hours) at reflux. MTBE (lOOmL) was added to the solution at 65°C and a solid product precipitated. A second portion of MTBE (lOOmL) was then added and the resulting slurry was cooled to room temperature over approximately 3 hours. The slurry was filtered and the wet cake was washed with MTBE (3x 100mL). The white wet cake was dried in a vacuum oven at ~50°C overnight to afford a white solid product (14.7g, 94.8% yield).

The exchange reactions were performed by dissolving the fatty acid salt with glycopyrronium bromide in Me-THF/water, stirring for 5.0 h at room temperature, then

WO 2016/204998

PCT/US2016/035967 separating the phases and washing the organic phase with water. The HPLC analytical results of both Me-THF and water layers are shown in Figs. 3 (potassium stearate) and 4 (potassium palmitate). The exchange reactions with potassium stearate and potassium palmitate did not provide the desired products free of the Acid A impurity. After concentration of the extracts, both reactions gave sticky oily residues that were difficult to isolate. Neither offered an advantage over the results from potassium laurate.

Stabilization of Exchange Reactions by Adding Excess free Fatty Acid

In an attempt to stabilize the glycopyrronium laurate product and avoid the decomposition of glycopyrronium to Acid A, the Me-THF/water reaction system was tested with excess free lauric acid (1.1 molar equivalent excess relative to glycopyrronium bromide). Two experiments with and without excess free lauric acid were conducted by mixing two solutions of glycopyrronium bromide (1.04g, 1.0 eq.) in water (5mL) and potassium laurate (1.1 eq.) in water (5mL) with Me-THF (20mL) with and without excess lauric acid (1.1 eq.) at room temperature overnight. For each reaction, both layers were checked by HPLC with the results of the HPLC for the exchange reaction with excess free lauric acid summarized in Table 16 below. The reaction with excess free lauric acid showed no formation of the by-product, Acid A, after mixing overnight but the experiment without excess free lauric acid, had Acid A in both layers.

Table 16

Experiment	HPLC Area Ratio of “Bromide’7”Glycopyrronium”
Me-THL Layer	Aqueous Layer
Without Lauric Acid [1]	9.1/90.9111	80.8/19.2121
With Excess free Lauric Acid (1-1 eq.)	7.4/92.6131	80.6/19.4

[1] 5.5 area% of Acid A was observed in the Me-THF layer.

[2] 2.2 area% of Acid A was observed in aqueous layer.

[3] The lauric acid peak co-eluted in glycopyrronium peak.

Adding 1.1 molar equivalents of excess free lauric acid into a preparation with potassium laurate and glycopyrronium bromide in the water/Me-THF system stabilized the reaction mixture.

WO 2016/204998

PCT/US2016/035967

Bi-phasic reaction conditions enabled the desired exchange reactions between glycopyrronium bromide and alkali and alkaline earth metal salts of fatty acids. Favorable partitioning of the glycopyrronium moiety into the organic phase (along with the fatty acids) and partitioning of the bromide into the aqueous phase was achieved with water and methyl tetrahydrofuran. The glycopyrronium fatty acid salts were unstable with respect to hydrolysis under the reaction conditions and were unstable as isolated oily products. The formation of the impurity, Acid A, was noted over time. An excess of the free fatty acid in the reaction mixture stabilized the glycopyrronium fatty acid salt and reduced the formation of the impurity, Acid A.

Reaction with Methyl Carbonates

In the prior art, dimethyl carbonate chemistry has been mostly employed in the surfactants and detergents industry as well as in the manufacture of ionic liquids with counterions that are different from halides.

The reaction of glycopyrrolate base with methyl carbonate and subsequent 15 treatment with fatty acids failed to provide any of the desired product(s) and produced only decomposition products from the glycopyrrolate base.

One example of the chemistry involved in this process is shown below:

Example

WO 2016/204998

PCT/US2016/035967

..•γ

GiycopyrroSste Base

Solvents

Gfycepy irons urn Methyl C a ft? on si $ Salt χ

Η · N, C-AbrCOr *CO-> -ο V,/ ca - Gfycopyrronium Laura to

The synthesis approach proceeding through methyl carbonates was unsuccessful, due to the instability of the starting glycopyrrolate base (compound 3 in Table 17) and its quaternization reaction products (glycopyrronium methyl carbonates, compound 4 in

Table 17) under the high temperature and pressure conditions required for a successful quaternization reaction with dimethyl carbonate.

Since a clean synthesis of the required methyl carbonated salt of glycopyrronium could not be accomplished, this approach was discontinued and subsequent efforts focused on a direct salt metathesis using glycopyrronium bromide.

Tertiary amines can be alkylated with dimethyl carbonate to generate the corresponding quaternary methyl ammonium carbonate salts in good to high yields. Subsequent reaction with a proton source leads to a clean ion metathesis reaction via decomposition of the methyl carbonate counterion into CO₂ and methanol, which are readily removed from the reaction mixture to yield products with high purity.

WO 2016/204998

PCT/US2016/035967

Glycopyrrolate Base

O

och₃

Solvent 80- 150 °C up to 100 psi

Glycopyrronium Methyl Carbonate Salt

Glycopyrronium Fatty Acid Salts

Table 17 shows the sequence of steps for synthesis of glycopyrronium salts of fatty acids via methyl carbonates from glycopyrrolate base.

The alkylation reaction requires high temperatures (typically 80-150 °C, depending of the reactivity of the tertiary amine) and is run under pressure, as dimethyl carbonate has a boiling point of 90 °C. Heterocyclic tertiary amines are reported to react faster and at relatively lower temperatures than the sterically crowded aliphatic tertiary amines. For example, Mori et al., US Patent No. 4,892,944, issued January 9, 1990 and incorporated herein by reference, report that N-methyl pyrrolidine (the closest structural motif to the glycopyrrolate base 3) can be quaternized at 120 °C to afford the corresponding methyl carbonate salt in 97% isolated yield after a 6 hour reaction. Bicyclic amines bearing a nitrogen atom at the bridgehead, which are known to be even more nucleophilic, react readily at atmospheric pressure under refluxing conditions (80 - 90 °C) to afford the methyl carbonates in high yields, (see Friesen et al., US Patent Publication 2012/0321969, published December 20, 2012, herein incorporated by reference).

The quatemization reaction was evaluated in neat dimethyl carbonate, and in two non-nucleophilic solvents, t-amyl alcohol and dimethyl acetamide, which were expected to

WO 2016/204998

PCT/US2016/035967 be unreactive toward the starting glycopyrrolate base. The glycopyrrolate base, compound

3, was custom made by Shanghai Chempartner (China) based on a published procedure (see Allmendinger et al., Carry Over of Impurities: A Detailed Exemplification for

Glycopyrrolate (NVA237), Organic Process Research & Development 2012,16, 1754 5 1769, herein incorporated by reference).

Four reactions of glycopyrrolate base (structure 3 in Table 3B) with dimethyl carbonate (shown in Table 18) were attempted. For the general procedure, glycopyrrolate base (structure 3 in Table 3B) was charged into a Fisher-Porter pressure bottle followed by methyl carbonate and solvent. The solution was purged with nitrogen three times under slight vacuum and then the reactor was sealed. The reaction mixture was stirred and heated with oil bath. The detailed reaction conditions and results are summarized in Fig.

5.

Glycopyrrolate Base

och₃

Solvents

Glycopyrronium Methyl Carbonate Salt

Glycopyrronium Fatty Acid Salts

In all four experiments (see Fig. 5), the glycopyrrolate base 3 was observed to decompose under the reaction conditions, yielding the cyclopentyl mandelic acid derivative (“Acid A”) and the corresponding methyl ester (Table 5). The analytical results showed that no desired product was observed in any of the reaction mixtures.

WO 2016/204998

PCT/US2016/035967

Due to the instability of the starting glycopyrrolate base, the salt metathesis approach via the glycopyrronium methyl carbonate salt was abandoned.

Preparation of HPLC Calibration Curves for Glycopyrrolate base and Glycopyrronium

Bromide

Based on analytical result reports for the raw materials, an HPLC method was setup and three HPLC calibration curves for glycopyrrolate base (structure 3 in Table 3B) as well as glycopyrronium bromide (structure 2 in Table 3B) were prepared. These are summarized in Table 19 and Figs. 1-2 and 6.

Table 19

Column	Waters XBridge C18 150x4.6 mm; 3.5pm
Flow Rate	1.0 mL/min
Column Temperature	40°C
Detection Wavelength	214 nm
Mobile Phase A	0.01M of NH4HCO3 Water
Mobile Phase B	Acetonitrile
Gradient	0 to 10 min MP B from 5% to 100% 10 to 15 min MP B at 100% 15.1 to 20 min MPBat5%
Glycopyrrolate Retention Time	~9.5 minutes
Glycopyrronium bromide Retention Time [Note]	Peak#l at ~1.5 minutes Peak #2 at -7.2 minutes

Note: Glycopyrronium bromide gave two distinct peaks in the chromatogram. Peak#l (RRT=0.20) corresponds to the “Bromide” (Br ion) and Peak #2 (RRT=1.00) corresponds to the “Glycopyrronium” moiety.

Further Synthesis and Analysis

Analytical methods were also developed for the glycopyrronium fatty acid salts.

These methods appropriately and accurately assess the quality of glycopyrronium fatty acid salt samples. A more refined process for the general synthesis of glycopyrronium fatty acid salts has also been developed.

WO 2016/204998

PCT/US2016/035967

The analysis includes both improvement in the synthesis of the glycopyrronium fatty acid salts and the development of methods to analyze those salts.

The synthesis process was improved and simplified by generating the required sodium or potassium salt of the fatty acids in situ (via treatment of the fatty acids with sodium or potassium hydroxide) rather than using sodium or potassium salts of the fatty acids prepared in a separate step. Substantial efforts were expended in defining a suitable amount of excess free fatty acid that would stabilize the product by minimizing its hydrolysis, while also facilitating the extraction process by minimizing the formation of emulsions, which were quite severe in some instances. Most of the development work in this second phase of the project was done with stearic acid. Also, in addition to lauric, palmitic and stearic acid, which are all saturated fatty acids, linoleic acid (C-18 unsaturated fatty acid; with cis,cis-9,12 double bonds) was also successfully converted to a glycopyrronium salt, further demonstrating the generality of the method.

The analytical methods development addressed the challenge stemming from the nature of the components from the reaction mixture, which required the use of complementary analytical methods to assess the salt metathesis process and the quality of the product. With the methods that have been developed, the salt metathesis process can now be well monitored and the quality of the fatty acid salts obtained reliably assessed.

The process that has been developed may be optimized to permit large scale production.

Some of the analytical method applications include, but are not limited to, assessment of the salt exchange effectiveness, to ensure that the reactions are pushed to >95% conversion by determining optimal number and amount of aqueous washes to remove inorganic bromide salts (byproducts), the determination of the optimal amount of excess free fatty acid (evaluating a suitable range of molar equivalents) needed to stabilize the active pharmaceutical ingredient (API) and the improvement of the isolation and purification of the reaction product (for example through precipitation with suitable solvent/antisolvent combinations) by evaluating the purity of the reaction products. In some preferred embodiments, the optimal number of aqueous washes is 3-4 washes. In other embodiments, the preferred number of washes is at least three washes.

WO 2016/204998

PCT/US2016/035967

Analytical method development also preferably includes methods for quantitation of chromophoric starting materials, products and degradants, methods for quantitation of weakly- or non-chromophoric starting materials and degradants (fatty acids and derived salts, dimethylhydroxypyrrolidinium degradants), methods for quantitation of the bromide ion in the starting material and the product, and methods for quantitation of the potassium (or sodium) ion in the product.

Glycopyrronium Fatty Acid salts (GPFA) were successfully prepared from glycopyrronium bromide and fatty acid salts utilizing a bi-phasic (organic / aqueous) system to selectively partition the inorganic salts into the aqueous phase and isolating the lipophilic organic salts (glycopyrronium fatty acid salts) from the separated organic phase. However, as discussed above, the reaction mixture was unstable to hydrolysis and also the fatty acid salts were unstable in methanol solution (transesterification by-product formation). The isolated glycopyrronium fatty acid salts (GPFA) also proved to be inherently unstable with increasing levels of the degradant, ’’Acid A” (CAS 427-49-6,acyclopentyl-a-hydroxy-benzene acetic acid), forming over time. An excess of free fatty acid stabilized the reaction mixture and significantly reduced the rate of hydrolysis (limiting the formation of “Acid A”). It was also challenging to accurately quantitate all components from the reaction mixtures to assess the synthesis efficiency.

Further experiments explored the isolation of glycopyrronium fatty acid salts (GPFA) with an excess of free fatty acids to confirm improved stability of the isolated products and also to develop suitable analytical methods to characterize the isolated materials. These experiments included preparing the samples for development of analytic methods, developing analytical methods to characterize the glycopyrronium fatty acid products, defining a suitable preparatory procedure and characterization of the isolated products, and preparing a range of GPFA samples. These efforts were successful, resulting in a suitable preparatory procedure and analytical methods to analyze and characterize the isolated products.

WO 2016/204998

PCT/US2016/035967

The developed analytical methods are reliable and simple enough to be tailored to the individual glycopyrronium salts of fatty acids (with anticipated minor modification of the GC method conditions for each fatty acid utilized).

As discussed above, only one approach (counterion exchange, or salt metathesis via a selective partitioning in a biphasic aqueous/organic solvent system) proved effective in the preparation of glycopyrronium fatty acid salts. Attempted synthesis from glycopyrrolate base via quaternization with methyl carbonate at elevated temperatures and subsequent treatment of the expected quaternary methylammonium carbonate salts with fatty acids was unsuccessful due to decomposition of glycopyrrolate base under the methylation conditions. Salt exchange in organic solvents under anhydrous conditions failed due to poor solubility of the glycopyrronium bromide and the fatty acid salt in the organic solvents that were evaluated (acetone and dichloromethane). The observed hydrolytic instability of the glycopyrronium moiety precluded consideration of aqueous ion exchange with ion exchange resins.

Even with the selective bi-phasic partitioning approach, severe emulsions and poor partitioning were observed for most systems tested. Selective partitioning of bromide into the aqueous phase and glycopyrronium into the organic phase was not easily achieved. Among the solvent systems screened (2-Me-THF, MTBE, IP Ac, Toluene, MIBK, and DCM), only 2-Me-THF provided adequate selectivity for the desired partitioning. Also, hydrolytic instability was noted over the course of the work-up and isolation with the formation of “Acid A” over time in both extracted solutions and isolated products.

Isolated glycopyrronium fatty acid mixtures ranged in consistency from viscous oils to pastes to hard waxy solids.

The initial synthesis experiments tested lauric acid (C-12 chain), stearic acid (C-18 chain) and palmitic acid (C-15 chain). The further synthesis and analysis experiments tested stearic acid, lauric acid, palmitic acid, and linoleic acid (C-18 polyunsaturated, cis, cis-9,12-Octadecadienoic acid).

WO 2016/204998

PCT/US2016/035967

The method synthesized glycopyrronium fatty acid salts by reacting glycopyrronium bromide with a fatty acid salt in a biphasic reaction mixture and including at least a 0.2 molar equivalent of excess free fatty acid to the reaction mixture, in preferred embodiments, at least a 0.2 molar equivalent of excess free fatty acid is added to the reaction mixture to form a glycopyrronium fatty acid salt. In some preferred embodiments, between 0.2 and 1.2 molar equivalent of excess free fatty acid is added to the reaction mixture. In another preferred embodiment, at least 0.6 molar equivalent of excess free fatty acid is added to the reaction mixture. In yet another preferred embodiment, between 0.6 and 1.2 molar equivalent of excess free fatty acid is added to the mixture to form a glycopyrronium fatty acid salt. In another embodiment, approximately 1.2 molar equivalent of excess free fatty acid is added to the reaction mixture. In another embodiment, at least 1.1 molar equivalent of excess free fatty acid is added to the reaction mixture.

As described herein, excess free fatty acid is relative to the glycopyrronium bromide and fatty acid salt used. For example, for the lauric acid reaction, it is excess free lauric acid relative to the glycopyrronium bromide and potassium laurate used.

The excess free fatty acid stabilizes the reaction mixture. The larger excesses of free fatty acid (0.6-1.2 molar excess) improved the phase separations, improved stability of the organic extract solutions, and improved stability of the isolated products.

In preferred embodiments, the required sodium, potassium or calcium salt of the fatty acids is generated in-situ via treatment of the fatty acids with a metal hydroxide rather than using sodium, potassium, or calcium. In this step, a fatty acid is mixed with the metal hydroxide in the biphasic reaction mixture (preferably water and 2-methyltetrahydrofuran) until all solids are dissolved to form the fatty acid salt. The metal hydroxide is preferably either an alkali metal hydroxide or an alkaline earth metal hydroxide (e.g.- sodium, potassium or calcium hydroxide).

The method also preferably includes adding glycopyrronium bromide and mixing until the solids dissolve. The upper organic phase is retained, while the lower aqueous

WO 2016/204998

PCT/US2016/035967 phase of the reaction mixture is removed. The upper organic phase is then washed with water. The steps of retaining the upper organic phase and removing the lower aqueous phase are preferably repeated. In preferred embodiments, these two steps are repeated at least 3 times to increase the purity of the glycopyrronium fatty acid salt. In other preferred embodiments, these two steps are repeated 3-4 times.

Vacuum distillation is preferably performed on the upper organic phase and new 2methyl-tetrahydrofuran is added. The vacuum distillation step is preferably repeated at least once (for a total of two times), to remove trace amounts of residual water, until no distillate is observed. In preferred embodiments, a non-polar hydrocarbon solvent is added to precipitate insoluble components, which are filtered off. The filtrate is then further distilled under reduced pressure to obtain a solid. In preferred embodiments, the non-polar hydrocarbon solvent is n-heptane or a mixture of heptane isomers. In alternative embodiments, n-hexane, isooctane or petroleum ether could be used instead of heptanes.

Analytical methods were needed that would be sufficient to characterize the composition and purity of the isolated products and enable further development. No single method was suitable for all components and complementary (orthogonal) methods were required to adequately characterize the isolated products.

Development of a purity and assay method was hindered by the lack of chromophores for the fatty acid components and any 3-hydroxy-1,1-dimethyl pyrrolidinium degradants. There was also a lack of retention/separation between bromide and 3-hydroxy-1,1-dimethyl pyrrolidinium degradant by HPLC. Challenging solubility properties of the glycopyrrolate mixtures, sample precipitation, column plugging, and accelerated loss of column performance also hindered method development. Ultimately, these challenges were overcome by development of the methods described below for the characterization of the glycopyrronium fatty acid salts.

WO 2016/204998

PCT/US2016/035967

With the challenges encountered, the preparation of the target glycopyrronium fatty acid salts was not trivial and required inventiveness to achieve the preparation of stable, well characterized products suitable for further development as API products.

Methods for preparing glycopyrronium fatty acid salts

Bi-phasic reaction conditions enable the desired exchange reactions between glycopyrronium bromide and sodium and potassium salts of fatty acids. Favorable partitioning of the glycopyrronium moiety into the organic phase (along with the fatty acids) and partitioning of the bromide into the aqueous phase was achieved with water and 2-methyl tetrahydrofuran. The glycopyrronium fatty acid salts were unstable to hydrolysis under the reaction conditions and as isolated oily products in the absence of an excess amount of free fatty acid. The formation of the impurity, “Acid A”, was noted over time. An excess of the free fatty acid in the reaction mixture stabilized the glycopyrronium fatty acid salt and reduced the formation of the impurity, “Acid A”. One reason for the stabilization may be an aggregation between the glycopyrronium fatty acid salt and the excess free fatty acid, with the hydrophobic fatty acid chain shielding the ester linkage and protecting it from access by water. The isolation of a well-defined product, though challenging, was achieved.

The preparatory procedure was further defined and the isolated products characterized using analytic methods. In addition, a range of glycopyrronium fatty acid samples were prepared, including glycopyrronium laurate, glycopyrronium stearate, glycopyrronium palmitate and glycopyrronium linoleate.

Initially, alkali metal fatty acid salts were prepared and isolated prior to use in preparing the glycopyrronium fatty acid mixtures. A simpler preparative procedure for the glycopyrronium fatty acid mixtures without isolation of fatty acid metal salts is preferred. Since the fatty acid metal salts, excess free fatty acid, and glycopyrronium bromide were mixed in a 2-Me-THF/water mixture, a modified procedure prepared the fatty acid metal salt in solution. The target amounts of fatty acid and metal hydroxide readily dissolved in

WO 2016/204998

PCT/US2016/035967 a mixture of water and 2-Me-THF to directly provide the desired solution of fatty acid salt and excess free fatty acid.

This approach is detailed in the following example. Stearic acid (6.26 g, 22 mmol, 2.2 eq.) and potassium hydroxide (0.726 g, 11 mmol, 1.1 eq.) were dissolved in a mixture of water (50 mF) and 2-Me-THF (50 mF). The molar excess of the free stearic acid in this example was 1.2 molar equivalent of excess free stearic acid (relative to the glycopyrronium bromide) and a 1.1 molar equivalent of excess free stearic acid relative to potassium hydroxide, assuming the purity of the potassium hydroxide is 100% (which is rarely the case). Mechanical stirring provided complete dissolution over approximately 30 minutes. Without stirring, the mixture settled into two clear phases with little to no emulsion remaining after approximately 30 minutes. Glycopyrronium bromide (3.98 g, 10 mmol, 1.0 eq.) was added, mixed, and dissolved (immediate complete dissolution). The mixing was stopped and phase separation was complete in less than 30 minutes. The lower aqueous phase (pH=7) was removed and the upper organic phase was washed three times with 20 mL of water. Each phase separation required less than 1 hour. After removing the last water wash, the rich organic phase was vacuum concentrated to mushy paste at 20-25°C / 25-30 Torr. 2-Me-THF (30 mL) was added and the vacuum concentration was repeated. Heptane (50 mL) was added resulting in a thin slurry. Dissolution in heptane confirmed the lipophilic nature of the glycopyrronium stearate since both glycopyrronium bromide and potassium stearate are insoluble in heptane.

When the heptane solution was chilled in an ice bath, a thick slurry formed but this became a thin slurry on re-warming to ambient temperature. After cooling in an ice bath for approximately 30 minutes, the chilled slurry was filtered on a glass-frit funnel but filtration was very slow and the contents warmed to room temperature during filtration. The isolated solids, confirmed to be stearic acid, were sucked dry under a nitrogen blanket and the dried solids weighed 0.9 g. The filtrate was vacuum concentrated at 20-25°C / 2530 Torr to an oily paste. The filtrate was confirmed to be glycopyrronium stearate and excess free stearic acid.

These samples were utilized for analytical method development. During the course of the analytical method development, it was observed that the glycopyrronium

WO 2016/204998

PCT/US2016/035967 stearate sample (EE-008-001-3B) was significantly more stable than prior samples prepared during the initial feasibility work phase (much slower rate of “Acid A” formation). However, the sample was still observed to slowly degrade over time, and by the time the method development was completed (~ 4 months), and the sample was analyzed with the final methods, formation of “Acid A” was noted (2.44% w/w, 9.29% AUC HPLC Acid A relative to glycopyrronium content).

Lurther experiments tested variations on the excess of free stearic acid used to prepare the glycopyrronium stearate. These experiments used the same general procedure as discussed in the glycopyrronium stearate example (EE-008-001-3B) above. More specifically, different molar equivalents of stearic acid were tested, from 1.2 to 2.0 molar equivalents (from 0.2 to 1.0 excess molar equivalents), in 0.2 molar equivalent increments. The different variations of free fatty acid excess tested are shown in Table 20. Samples were waxy solids from Methyl-THE/GPBr/Stearic Acid/Potassium Stearate mixtures after three water washes.

Table 20

Sample ID	Description
EE-008-004-A	From 1.0 eq. GPBr, 1.1 eq. KOH, 1.2 eq. Stearic Acid
EE-008-004-B	From 1.0 eq. GPBr, 1.1 eq. KOH, 1.4 eq. Stearic Acid
EE-008-004-C	From 1.0 eq. GPBr, 1.1 eq. KOH, 1.6 eq. Stearic Acid
EE-008-004-D	From 1.0 eq. GPBr, 1.1 eq. KOH, 1.8 eq. Stearic Acid
EE-008-004-E	From 1.0 eq. GPBr, 1.1 eq. KOH, 2.0 eq. Stearic Acid

Samples A and B had difficult (very slow) phase splits for the water washes after the initial phase split of the reaction mixture. In contrast, all phase splits for samples C, D, and E were done within approximately 1 hour. In contrast to the earlier preparation of EE008-001-3B, with 2.2 molar equivalent of stearic acid, no filterable solids were observed from the final heptane dissolution; only some separation of an oily phase was noted and this was retained during isolation. The samples were analyzed and the results of the comparative analysis from variable fatty acid input are summarized in Table 21. Sample EE-008-001-3B was included for comparison.

Table 21

WO 2016/204998

PCT/US2016/035967

Sample ID	Stearic Acid Input	Wt.% K+	Wt.% Br	wt% GP	Wt.% Stearate	Wt.% Other	GP HPLC AUC	Molar ratio Stearate / GP
EE-008- 004-A	1.2 eq	0.24	0.036	32.53	49.42	18.05	82.8%	1.7
EE-008- 004-B	1.4 eq	0.27	0.058	21.82	5.85	72.33	96.4%	0.3
EE-008- 004-C	1.6 eq	0.25	0.020	24.16	10.1	65.73	98.0%	0.47
EE-008- 004-D	1.8 eq	0.32	0.006	34.65	32.78	32.57	96.1%	1.05
EE-008- 004-E	2.0 eq	0.28	0.006	29.30	24.29	46.41	97.1%	0.93
EE-008- 001-3B	2.2 eq	<0.1	<0.1	38.39	51.45	10.16	94.3%	1.5

While an effective purge of bromide was demonstrated and residual potassium was low, there was no clear trend in the results for glycopyrronium or stearate content across the samples. Since using less than 2.2 molar equivalent of stearic acid (relative to glycopyrronium) showed no advantage, the next experiment repeated the conditions of EE-008-001-3B (2.2 eq. stearic acid).

The sample that resulted from the repeat preparation of glycopyrronium stearate using 2.2 eq. stearic acid, EE-008-008, was analyzed and compared to EE-008-001-3B. The results are shown in Table 22.

Table 22

	Sample EE-008-008 (new)	Sample EE-008-001-3B (retest)
GP % w/w HPLC	31.4%	31.7%
GP % AUC Purity	95.89%	87.33%
“Acid A” % w/w	0.13%	2.44%
“Acid A” AUC %	0.50%	9.29%
Stearic Acid w/w% GC	74.1%	67.6%
Potassium ΉνΝο IC	1.40%	0.60%
Bromide ΉνΝο IC	0.008%	0.002%
GP/Stearate molar ratio NMR	1 : 2.35
GP w/w% by H1 NMR	32.2%

WO 2016/204998

PCT/US2016/035967

Stearate w/w% by H1 NMR

67.7%

While the level of the degradant “Acid A” was typically low in the freshly isolated product, this impurity increased over time (approximately 4 months) at ambient conditions.

For EE-008-008, the input was 3.98 g (10 mmol) glycopyrronium bromide (with 22 mmol stearic Acid and 11 mmol KOH). The output was a 5.35 g glycopyrronium stearate mixture. The mass recovery was 56.7%. The glycopyrronium recovery was 54.2% (by fflNMR) and 52.8% (by HPLC assay).

HPLC data for glycopyrronium stearate EE-008-008 is shown in Fig. 7. HPLC data for glycopyrronium stearate EE-008-001-3B is shown in Fig. 8. Peak 1, the largest peak, with 87.6482 percent area in EE-008-001-3B and 96.0360 percent area in EE-008008, is glycopyrronium stearate. The Acid A peak occurs at approximately 17.6 minutes retention time (peak 5 in Fig. 7 and peak 8 in Fig. 8).

Gas chromatography data for glycopyrronium stearate sample EE-008-008 is shown in Fig. 9. Gas chromatography data for glycopyrronium stearate sample EE-008001-3B is shown in Fig. 10. The far left peak in Fig. 9 and Fig. 10 is the solvent in the gas chromatography trace. This data shows the total amount of stearic acid (the peak at 19.827 in EE-008-008 and the peak at 19.836 in EE-008-001-3B) in the samples. The total amount of stearic acid includes the stearic acid in the glycopyrronium stearate, as well as any excess free stearic acid still present in the sample. The amount of excess free stearic acid can be calculated by determining how much stearic acid has been incorporated into the glycopyrronium stearate stoichiometrically.

All NMR chemical shifts are provided in ppm relative to tetramethylsilane (TMS). NMR data for glycopyrronium stearate sample EE-008-008 is shown in Fig. 11 and Figs. 12A-12C. *H NMR (400 MHz, CDC1₃) 0.88 (3H, t, CH₃-), 1.20 - 1.70 (30H, m, 15 -CH₂ groups from fatty acid chain and 8H, m, 4 -CH₂ groups from cyclopentyl ring), 2.0 - 2.25 (1H, m, -CH-C-N+), 2.24 (2H, t, -CH₂C=O), 2.7 - 2.9 (1H, m, -CH-C-N⁺ and 1H, m, CH group from cyclopentyl group), 2.93, 3.11, 3.33, 3.37 (6H, 4 sets of singlets, 2 -CH₃-N⁺;

WO 2016/204998

PCT/US2016/035967 chemical shift differences at the charged interface possibly due to different aggregation states of the fatty acid salt), 3.6 - 3.8 (2H, m, -CH2-N¹), 3.8 - 4.0 (1H, m, -CH-N⁺), 4.1 4.3 (1H, m, -CH-N⁺), 5.40 - 5.55 (1H, m, -CH-OH, methine proton), 7.2 - 7.7 (5H, m, aromatic protons/phenyl group).

NMR data for the Acid A by-product is shown in Figs. 13A-13C. *H NMR (400 MHz, CDCI3) 1.2 - 1.8 (8H, m, 4 CH2 groups from cyclopentyl ring), 2.9 - 3.0 (1H, m, CH group from cyclopentyl ring), 6.2 (1H, br s, OH/tertiary alcohol), 7.2 - 7.7 (5H, m, aromatic protons/phenyl group).

NMR data for the glycopyrronium Hydrolysis By-Product, Quaternary amino alcohol (QAA) is shown in Figs. 14A-14D. Ή NMR (400 MHz, D₂O) 2.1 - 2.3 (1H, m, CH- C-N⁺), 2.5 - 2.7 (1H, m, -CH-C-N⁺), 3.2 (3H, s, CH3-N⁺), 3.3 (3H, s, CH3-N⁺), 3.5 3.7 (2H, m, -CH2-N⁺), 3.7 - 3.9 (2H, m, -CH2-N⁺), 4.75 - 4.85 (1H, m, -CH-OH, methine proton).

When the glycopyrronium stearate hydrolyzes, it results in Acid A and QAA. QAA was very difficult to analyze using HPLC, because it does not have a strong chromophore and comes off very early in the HPLC analysis.

The loss of glycopyrronium fatty acid salts to the aqueous washes was higher than desired but the procedure is workable to prepare samples. Some methods to improve salt recovery include, but are not limited to, performing a back extraction, salting out, or choosing alternative extraction solvents. Both the glycopyrronium moiety and stearic acid are lost to the aqueous washes but the ratio of stearic acid to glycopyrronium in the isolated product is enriched compared to the input levels (2.2:1 as input, 2.35: 1 isolated) The final ratio may be further adjusted with adjustments in the free fatty acid input.

There was an effective purge of bromide during work up and isolation. More specifically, the work up and isolation procedure consistently provided very low levels (<0.1 %) of bromide. This may permit optimization of the washes for enhanced glycopyrronium recovery in the process while still providing control of residual bromide to acceptable levels.

WO 2016/204998

PCT/US2016/035967

The resulting glycopyrronium stearate sample was characterized by HPLC and GC and consistent with NMR results. Confirmed results by separate, complementary analytical methods validated the results. The sample stability was improved over the initial studies. The samples isolated with higher excess free fatty acid have improved stability with degradation observed only after months at ambient conditions. The excess free fatty acid appears to act as a buffer, to slow down degradation of the desired fatty acid salts.

Preparation of Glycopyrronium Fatty Acid Salts

As a more general description for preparing glycopyrronium fatty acid salts, the calculated target amount of fatty acid (FA) is mixed with water (approximately 8 mF/g FA input), 2-methyl-tetrahydrofuran (2-Me-THF, approximately 8 mF/g FA input), and metal hydroxide such as an alkali metal hydroxide or an alkaline earth metal hydroxide (e.g.potassium hydroxide KOH, 1.1 molar equiv). In preferred embodiments, the calculated target amount is 2.2 molar equivalents (a 1.2 molar equivalent excess). In other embodiments, the calculated amount is greater than or equal to 1.2 molar equivalents (0.2 molar equivalent excess). The mixture is stirred until all solids are dissolved.

Glycopyrronium bromide (“GPBr”, 1.0 molar equiv.) is added and mixed until the solids are dissolved. Mixing is stopped and the phases are allowed to separate. The lower aqueous phase (pH approximately 7) is removed and the upper organic phase is retained. The upper organic phase is washed three times with water (each wash ~3.2 mF/g fatty acid input). Each lower aqueous phase (pH ~7) is removed and the upper organic phase is retained.

The rich upper organic phase is concentrated by vacuum distillation using minimal heating (20-25°C / 25-30 Torr) to obtain a mushy paste. Fresh 2-Me-THF (approximately 4.8 mF/g FA input) is added and vacuum concentration is conducted (20-25°C / 25-30 Torr) until no further distillate is observed.

Heptane (approximately 8 mF/g fatty acid input) is added and the mixture is stirred until most of the solids dissolve. Some of the excess free fatty acid may remain as a thin

WO 2016/204998

PCT/US2016/035967 slurry which is passed through a filter to remove the remaining solids. The rich filtered heptane solution is concentrated by vacuum distillation using minimal heating (20-25 °C /

25-30 Torr) to obtain a waxy solid.

Additional fatty acids were used to prepare glycopyrronium fatty acid salts using 5 the methods described herein. More specifically, lauric acid (sample PSG-008-002), palmitic acid (sample PSG-008-204) and linoleic acid (sample PSG-008-206) were used as starting materials, in addition to the stearic acid (sample EE-008-008) samples discussed above. The results are shown in Tables 23A and 23B. The theoretical maximum recovery was calculated as total recovery of glycopyrronium fatty acid and excess free fatty acid.

Table 23A:

Sample Name	Lot#	File	Sample wt (mg)	GP PA (peak area)	Assay	Average Assay
GP Laurate	PSG-008- 202	09	33.86	1469.4	34.544%	34.703%
10	33.80	1480.3	34.862%
GP Palmitate	PSG-008- 204	11	36.74	1372.6	29.739%	29.934%
12	36.03	1363.7	30.129%
GP Linoleate	PSG-008- 206	13	37.99	1163.5	24.379%	26.749%
14	38.33	1402.1	29.118%

Table 23B

Sample	Theoretical Maximum Recovery*	Prep Output	Mass Yield	GP Wt% HPLC	Comments
EE-008-008 Stearic Acid	9.44 g	5.35 g	56.67%	31.4%	Waxy Solid
PSG-008-202 Lauric Acid	7.58 g	6.08 g	80.2%	34.7%	Oily Solid
PSG-008-204 Palmitic Acid	8-82 g	3.00 g	34.0%	29.9%	Oily Solid
PSG-008-206 Linoleic Acid	9.34 g	9.32 g	99.8%	26.7%	Oily Solid

*Calculated as total recovery of GPFA and excess free FA

The HPLC data for glycopyrronium stearate is shown in Figs. 7 and 8. The HPLC data for glycopyrronium laurate is shown in Figs. 15A-15D. Peak 1 indicates

WO 2016/204998

PCT/US2016/035967 glycopyrronium laurate, and includes an area % of 90.8652 and 90.8662, respectively in the two runs. Peak 7 indicates the by-product Acid A, which includes a 5.7508% area and a 5.758% area in the two runs. The other minor peaks are unidentified impurities or artifacts.

The HPLC data for glycopyrronium palmitate is shown in Figs. 16A-16D. Peak 1 indicates glycopyrronium palmitate, and includes an area % of 91.2525 and 91.3890, respectively in the two runs. Peak 5 indicates the by-product Acid A, which includes a 4.1180% area and a 4.0100% area, respectively in the two runs. The other minor peaks are unidentified impurities or artifacts.

The HPLC data for glycopyrronium linoleate is shown in Figures 17A-17D. Peak 1 indicates glycopyrronium linoleate, and includes an area % of 39.7889 and 39.5620, respectively for the two runs. Peak 4 indicates the by-product Acid A in the first run (Figs. 17A-17B) and peak 5 indicates the by-product Acid A in the second run (Figs. 17C-17D), and includes a 2.4126% area and 2.4171% area, respectively for the two runs. Peak 15 (Figs. 17A-17B and peak 17 (Figs. 17C and 17D) actually had the highest area percentage (55.1974 and 55.4318, respectively). These peaks are consistent with the excess free linoleic acid used in the reaction, which was detected in this HPLC due to the conjugated C=C double bond chromophore. Of the fatty acids tested, only the linoleic acid shows a significant response with the UV detector. The other minor peaks are unidentified impurities or artifacts. A more accurate assessment of the amount of free linoleic acid can be carried out by gas chromatography, with a minor optimization of parameters used for the analysis of stearic acid.

As an analysis tool, HPLC was only able to identify the desired product (glycopyrronium stearate, glycopyrronium laurate, glycopyrronium palmitate and glycopyrronium linoleate) and Acid A in the samples. Given the complexity of the reagents and products, other methods were required to identify the fatty acids (gas chromatography), the desired products and by-products (NMR), and the residual bromide and potassium (ion chromatography) in the samples.

WO 2016/204998

PCT/US2016/035967

All NMR chemical shifts are provided in ppm relative to tetramethylsilane (TMS).NMR data for glycopyrronium bromide is shown in Fig. 18. ¹H NMR (400 MHz, DMSO-dg) 1.1-1.7 (8H, m, 4 CH2 groups from cyclopentyl group), 2.0 - 2.2 (1H, m, CH-C-N⁺), 2.6 - 2.8 (1H, m, -CH-C-N⁺), 2.8 - 3.0 (1H, m, CH group from cyclopentyl group), 3.1 (3H, s, CH3-N⁺), 3.2 (3H, s, CH3-N⁺), 3.45 - 3.65 (2H, m, -CH₂-N⁺), 3.65 3.85 (1H, m, -CH-N⁺), 3.85 - 3.95 (1H, m, -CH-N⁺), 7.20 - 7.65 (5H, m, aromatic protons/phenyl group). NMR data for glycopyrronium stearate is discussed above and is shown in Fig. 11 and 12A-12C.

NMR data for glycopyrronium laurate is shown in Fig. 19. ¹H NMR (400 MHz, CDC13) 0.87 (3H, t, CH3-), 1.20 - 1.70 (18H, m, 9 -CH2 groups from fatty acid chain and 8H, m, 4 -CH2 groups from cyclopentyl ring), 2.1 - 2.25 (1H, m, -CH-C-N+), 2.23 (2H, t, -CH2C=O), 2.7 - 3.0 (1H, m, -CH-C-N+ and 1H, m, CH group from cyclopentyl group), 2.95, 3.10, 3.28, 3.30 (6H, 4 sets of singlets, 2 -CH3-N+; chemical shift differences at the charged interface possibly due to different aggregation states of the fatty acid salt), 3.6 3.85 (2H, m, -CH2-N+ and 1H, m, -CH-N+), 3.95 - 4.1 (1H, m, -CH-N+), 5.4 - 5.5 (1H, m, -CH-OH, methine proton), 7.2 - 7.6 (5H, m, aromatic protons/phenyl group).

NMR data for glycopyrronium palmitate is shown in Fig. 20. ¹H NMR (400 MHz, CDC13) 0.88 (3H, t, CH3-), 1.20 - 1.70 (26H, m, 13 -CH2 groups from fatty acid chain and 8H, m, 4 -CH2 groups from cyclopentyl ring), 2.0 - 2.2 (1H, m, -CH-C-N+), 2.25 (2H, t, -CH2C=O), 2.7 - 2.95 (1H, m, -CH-C-N+ and 1H, m, CH group from cyclopentyl group), 2.93, 3.11, 3.34, 3.37 (6H, 4 sets of singlets, 2 -CH3-N+; chemical shift differences at the charged interface possibly due to different aggregation states of the fatty acid salt), 3.6 - 3.8 (2H, m, -CH2-N+), 3.8 - 4.0 (1H, m, -CH-N+), 4.1 - 4.25 (1H, m, CH-N+), 5.40 - 5.55 (1H, m, -CH-OH, methine proton), 7.2 - 7.6 (5H, m, aromatic protons/phenyl group).

NMR data for glycopyrronium linoleate is shown in Fig. 21. ¹H NMR (400 MHz, CDC13) 0.88 (3H, t, CH3-), 1.20 - 1.70 (16H, m, 8 -CH2 groups from fatty acid chain and 8H, m, 4 -CH2 groups from cyclopentyl ring), 2.0 - 2.1 (4H, m, allylic protons -CH2C=C), 2.0 - 2.2 (1H, m, -CH-C-N⁺), 2.2 (2H, t, -CH2C=O), 2.85 (2H, t, doubly allylic

WO 2016/204998

PCT/US2016/035967 protons C=C-CH₂-C=C), 2.7 - 2.9 (1H, m, -CH-C-N⁴ and 1H, m, CH group from cyclopentyl group), 2.91, 3.09, 3.29, 3.31 (6H, 4 sets of singlets, 2 -CH₃-N⁺; chemical shift differences at the charged interface possibly due to different aggregation states of the fatty acid salt), 3.6 - 3.9 (2H, m, -CH₂-N⁺ and 1H, m, -CH-N⁺), 4.0 - 4.15 (1H, m, -CH5 N⁺), 5.25 - 5.50 (2H, m, olefinic protons), 5.35 - 5.50 (1H, m, -CH-OH, methine proton),

5.6 - 5.75 (2H, m, olefinic protons), 7.2 - 7.6 (5H, m, aromatic protons/phenyl group)

NMR analysis could be effectively used to accurately assess the molar ratio of glycopyrronium to the total amount of fatty acid component (i.e. stoichiometrically bound fatty acid anion and free fatty acid). A comparison of H¹ NMR results for the glycopyrronium fatty acid salt samples is shown in Table 24. Interfering peaks were noted in the aromatic region for the stearic acid sample, but the glycopyrronium multiplets were sharp with no interference.

Table 24

Sample	Integration of Fatty Acid Methyl (3 H total)	Integration of GP Aromatic Peaks (5 H total)	Integration of GP Multiplets (4 H total) 3.5 - 4.2 ppm	Mole Ratio Fatty Acid/GP Calculated from FA Methyl & GP Multiplets
EE-008-008 Stearic Acid	1.41 (H = 0.47)	1.70 (H = 0.34)	0.82 (H = 0.205)	2.29 : 1
PSG-008-202 Lauric Acid	4.07 (H=1.3567)	2.65 (H=0.53)	1.99 (14=0.4975)	2.73 : 1
PSG-008-204 Palmitic Acid	2.54 (14=0.8467)	1.78 (H=0.356)	1.18 (14=0.295)	2.87 :1
PSG-008-206 Linoleic Acid	3.57 (H= 1.19)	2.66 (14=0.532)	1.97 (14=0.4925)	2.42 :1

The glycopyrronium HPLC assay (GP wt% Content) is applicable to a variety of fatty acid salts. While the HPLC method was developed with glycopyrronium stearate, no interfering peaks or other issues were observed in the analysis of glycopyrronium fatty acid salts prepared from lauric, palmitic and linoleic acids enabling quantitation of the glycopyrronium content in these mixtures.

WO 2016/204998

PCT/US2016/035967

Similarly, NMR effectively characterizes a variety of fatty acid salts using a number of different fatty acid substrates. As in the case of glycopyrronium Stearate, the terminal methyl group of the fatty acid chain (3 H) provided suitable integration for comparison with the glycopyrronium (GP) multiplets (4 H total, 3.5 - 4.2 ppm) for calculation of the mole ratio (GP:FA) present in the isolated materials for glycopyrronium fatty acid salts prepared from lauric, palmitic, and linoleic acids.

The glycopyrronium fatty acid salt samples were relatively stable. The level of the degradant “Acid A” was low in all samples at the initial isolation but stability over time (months) was only examined for glycopyrronium stearate. For glycopyrronium stearate, degradation was observed over ~4 months at ambient conditions.

The mass recovery was variable across the different fatty acids. The mass recovery varied from worse to much better across the glycopyrronium fatty acid samples (Palmitic < Stearic < Lauric < Linoleic). This highlights the opportunity for additional procedure optimization by further improving retention of glycopyrronium salt activity and recovery in the isolated glycopyrronium fatty acid salts. Some methods to improve salt recovery include, but are not limited to, performing a back extraction, salting out, or choosing alternative extraction solvents.

Extended drying was required to achieve a low level of residual n-heptane (drying until heptane undetectable by NMR), particularly with lauric and palmitic acids. Further optimization of drying conditions may improve the process. For example, the vacuum and/or temperature of the drying conditions could be adjusted to improve the process while avoiding degradation.

Similar mole ratios were noted for fatty acid :glycopyrronium (FA:GP) in the final products. All of the isolated glycopyrronium fatty acid salt mixtures showed some enrichment of the fatty acid (relative to glycopyrronium) compared to the input ratio. The isolated products all had more than the 2.2:1 FA:GP input ratio.

Fatty acids enjoy a GRAS (Generally Regarded as Safe) status, are widely used in drug formulations, and are found in many foods that are currently consumed. So, a large

WO 2016/204998

PCT/US2016/035967 excess of free fatty acid in the glycopyrronium fatty acids is not likely to pose a problem for the use of the glycopyrronium fatty acids in drug development.

The current preparative and analytical methods may be applied to other fatty acids, preferably fatty acids with at least eight carbon molecules. Some examples include, but are not limited to, arachidic acid, stearic acid, palmitic acid, oleic acid, erucic acid, arachidonic acid, lauric acid, capric acid, linoleic acid, linolenic acid, or myristic acid to make glycopyrronium fatty acid salts.

Analysis of products and degradants

The analytical methods needed to be sufficient to characterize composition and 10 purity as meeting regulatory requirements to enable further development. The lack of chromophores for fatty acid components and 3-hydroxy-1,1-dimethyl pyrrolidinium degradants required the development of alternatives to HPLC/UV. The lack of retention / separation of bromide and 3-hydroxy-1,1-dimethyl pyrrolidinium by HPLC required the development and use of separate, complementary methods to quantitate these components.

Challenging solubility properties of the active pharmaceutical ingredient mixture (sample precipitation, column plugging, accelerated loss of column performance) hindered method development but these issues were ultimately overcome. No single method was suitable for all components of analysis so complementary (orthogonal) methods were required and successfully developed.

Samples of a crude product mixture and each of the key hydrolysis by-products were prepared and analyzed along with samples of the starting materials to enable method development. The samples prepared are shown in Table 25.

Table 25

Sample ID	Description
EE-008-001-1	Glycopyrronium Bromide (GPBr), lot # 6908953CD Starting Material
EE-008-001-2	Stearic Acid, SAFC lot # 39397PJ Starting Material (used in excess)

WO 2016/204998

PCT/US2016/035967

EE-008 -001-3a	Filterable Solids from GPBr / Stearic Acid / Potassium Stearate (recovered portion of stearic acid)
EE-008-001-3b	Waxy Solids from GPBr / Stearic Acid / Potassium Stearate API mixture of GP Stearate and excess free Stearic acid
EE-008-001-4a	GP Hydrolysis By-Product, Organic Acid (“Acid A”) CAS 427-49-6, a-cyclopentyl-a-hydroxy-benzene acetic acid
EE-008-001-4b	GP Hydrolysis By-Product, Quaternary aminoalcohol (“QAA”) CAS 51052-74-5, 3-hydroxy-l,l-dimethylpyrrolidinium bromide mixture with NaBr

By-Product (Degradant) Sample Preparation

Hydrolysis of glycopyrronium bromide and isolation of the by-products was performed. A mixture of 3.98 g (10 mmol, 1.0 eq.) glycopyrronium bromide, 0.80 g (20 mmol, 2.0 eq.) sodium hydroxide, and 20 mL of water was stirred at room temperature overnight. The resulting solution was filtered through filter paper to remove traces of sticky solids and then acidified by the drop-wise addition of 4.78 g (28.35 mmol, 2.835 equiv.) of 48% Aq. hydrobromic acid. A thick white slurry formed which was then filtered and washed with ~5 mL of water. The isolated white solids were air dried giving

2.025 g of “Acid A” (CAS 427-49-6, α-cyclopentyl-a-hydroxy-benzene acetic acid). The combined filtrate and rinse was vacuum concentrated at 40°C / 9 Torr to provide 3.87 g of an oily paste containing the Quaternary Amino Alcohol By-Product (CAS 51052-74-5, 3hydroxy-1,1-dimethyl pyrrolidinium bromide) as well as NaBr in -0.95/1 w/w ratio.

Glycopyronnium Bromide mw 398.34

2) Acidify Aq. HBr room temp.

l)NaOH/Water room temp.

Acid By-Product Acid A mw 220

Vacuum Concentrate Acidic Filtrate

Θ

Br + NaBr

Filter and Water Wash White Crystalline Precipitate, -92% Yieled

Quaternary Amino Alcohol By-Product and NaBr mixture isolated as oily paste 0.95/1 w/w ratio

HPLC Methods

A unique HPLC method was developed to analyze the products of the reaction. Table 26 shows screening for HPLC method conditions.

WO 2016/204998

PCT/US2016/035967

Table 26

Columns Screened:	Mobile Phases Used
Phenomenex Luna Cl8(2) 4.6 x 150mm, 3μ Waters Atlantis T3, 4.6 x 150mm, 3μ SIELC Primesep 100, 4.6 x 150mm, 5μ Waters Symmetry Cl8, 4.6 x 150mm , 3.5μ Zorbax Eclipse XDB C8, 4.6 x 150mm, 5μ Zorbax Eclipse XDB C18, 4.6 x 150mm, 5μ Phenomenex Hydroxy RP, 4.6 x 250mm, 5μ	1st set: Mobile Phase A: 0.1% H3PO4 in DI H2O Mobile Phase B: 0.1% H3PO4 in ACN 2nd set: Mobile Phase A: 0.01M KH2PO4 pH 6.5 in DI H2O Mobile Phase B: 100% ACN
	3rd set: Mobile Phase A: 0.01% TFA in DI H2O Mobile Phase B: 0.01% TFA in ACN

In addition to the listed screening, the USP method for Glycopyrrolate was evaluated. This method used Kinetex Cl8, 4.6 x 100mm, 2.6μ, Mobile Phase A: 0.025M

KH₂PO₄ pH 2.5 with H3PO4 in DI H₂O, and Mobile Phase B: 100% CAN. As per the USP method, the diluent was 1:1 mobile phase A and B, at 0.5mg/mL sample concentration.

The first sample of glycopyrronium stearate (EE-008-001-3b, waxy solids from GPBr / Stearic Acid / Potassium Stearate, Table 25) seemed soluble in the USP diluent but methanol was used for consistency across all of the screened conditions and all samples were prepared at 0.5 mg/mL in MeOH.

With the USP Method, the baseline was not good (very irregular baseline). Also, numerous small peaks were observed on replicate injections. Precipitation on the column was suspected as the cause of the problem, so mobile phase mixtures were checked for precipitation. It was found that, at the high end of the gradient (15% mobile phase A and

85% mobile phase B), KH2PO4 buffer precipitated. Most of the C18 columns adequately retained glycopyrrolate (GP). The peak shape (GP) was found to be better at the lower pH (phosphoric acid).

From the screened conditions, the Agilent, ZORBAX Eclipse XDB-C18, 4.6 x 150mm, 5μηι, P/N: 993967-902 and the 0.1% H3PO4 in H2O/acetonitrile solvent system provided the best performance (baseline and peak shape). However, the degradant “QAA” (shown in Table 2, item 8, sample EE-008-001-4b) co-eluted with the bromide peak.

WO 2016/204998

PCT/US2016/035967

Despite extensive efforts, the retention and separation of the bromide and QAA peaks could not be improved.

Glycopyrronium and the “Acid A” degradant were readily resolved from other components and quantitated by HPLC and a weight percent assay for glycopyrronium content was developed. While it was desired to directly quantitate the QAA content by HPLC, it remained necessary to determine bromide by a separate method (IC) and then estimate the QAA content by HPLC using a calculation to subtract the bromide content. During the course of the HPLC method development, it was observed that “Acid A” was not observed in sample EE-008-001-3b for the first month of use.

One analytical method that was developed was a method for quantitation of chromophoric starting materials, products and degradants. The method used HPLC with photodiode array detection (PDA).

A glycopyrronium stearate HPLC method was developed, which provides a procedure for the determination of glycopyrronium or

3[(cyclopentylhydroxyphenylacetoxy]-l, 1-dimethyl pyrrolidinium (GP) stearate assay and impurity profile by HPLC.

The HPLC operating conditions for the Agilent, ZORBAX Eclipse XDB-C18, 4.6 x 150mm, 5μηι, P/N: 993967-902 column are shown in Table 27 and the mobile phase gradient is shown in Table 28. 0.1% of 0.25mg/mL GPBr, average S/N=l 1.2. Linearity from 25% to 120% of 0.25mg/mL GPBr is linear. The correlation coefficient= 1.000.

WO 2016/204998

PCT/US2016/035967

Table 27

Injection Volume	10 pL
Mobile Phase A	0.1% H3PO4in H2O
Mobile phase B	0.1% H3PO4in ACN
Gradient	See table below
Flow Rate	1 mL/min
Runtime	40 mins
Column Temperature	30C
UV wavelength	210 nm; For Diode Array detector, use 210nm Bw=8, reference 360nm, Bw=100

Table 28

Time (minutes)	A%	B%
0.0	90	10
5.0	90	10
25.0	5	95
35.0	5	95
35.1	90	10
40.0	90	10

Fig. 22 shows the glycopyrronium concentration versus glycopyrronium peak area.

Fig. 23 shows a blank chromatogram and Fig. 24 shows a resolution solution chromatogram. Method accuracy is shown in Table 29.

WO 2016/204998

PCT/US2016/035967

Table 29

Name	Recovery
80%-l	100.77%
80%-2	101.37%
100%-l	101.20%
100%-2	99.909%
120%-l	101.35%
120%-2	100.94%
Average	100.92%
Standard Deviation	0.5%

Ion Chromatography for Bromide Content

Another method quantitated the residual bromide ion in the product using ion chromatography (anion mode). The first rough check of the glycopyrronium stearate sample EE-008-001-3b (Table 10 p.19) showed <1% w/w bromide (limit test). A full calibration sequence and analysis was carried out. All results indicated effective removal of bromide. The final bromide assay method can check and control residual bromide ion to <0.1%.

This method was developed to provide a procedure for the determination of trace amounts of residual bromide ion in glycopyrronium stearate/ stearic acid mixtures using ion chromatography. The analytical method applies to the quantitation of residual bromide ion in glycopyrronium stearate/ stearic Acid mixtures. The method corresponds to 0.1% (1000 ppm) limit for bromide ion of a 0.5 mg/mL sample concentration.

WO 2016/204998

PCT/US2016/035967

The ion chromatographic parameters and conditions are shown in Table 30.

Table 30

Instrument	Parameter	Condition
Auto Sampler AS-DV	Sampler Flush Factor	10
Sampler Delay Volume	125 pL
Sampler Deliver Speed	4.0 mL/min
Sampler Deliver Sample	Full
RFIC	Data Collection Rate	5.0 Hz
Column Temperature	30° C
Cell Temperature	35° C
Eluent Generator Cone.	10 mM
Suppressor Current	10 mA
Flow Rate	0.5 mL/min
Injection Volume	10 pL
Run Time	21.0 min
Program Parameters	Stage	Time (min)	Command	Value
Equilibration	-3.000 -2.700	Pump ECD.Pump ECD Relay l.Closed	Duration = 138 seconds
Start Run	0.00	Pump_ECD.Channel_Pressure.AcqOn Pump_ECD. Autozero Pump_ECD.ECD_l. AcqOn Pump_ECD_Total.AcqON Pump_InjectValve.Inject Position	Duration - 30 seconds
Run Time	0.00		Duration = 18 minutes

Ion chromatography data for a bromide standard is shown in Fig. 25. Ion chromatography data for bromide in glycopyrronium stearate is shown in Fig. 26. The non-labeled peaks in Fig. 26 are the non-bromide charged species present in the sample matrix (artifacts, compare with bromide standard trace). As shown in Fig. 26, very little bromide remains in the sample, which indicates that the reaction was successful in producing the desired result, glycopyrronium stearate.

Ion Chromatography for Potassium Content

Yet another method quantitated the residual potassium (or sodium) ion in the product using ion chromatography (cation mode). The key challenges were the limited

WO 2016/204998

PCT/US2016/035967 solubility of the sample in a suitable diluent and the initial observation of interfering peaks near the potassium retention time. These were resolved and the method can detect potassium ion down to a limit of quantitation of 0.1% level for a 0.5 mg/mL API sample.

Ion chromatography data for potassium in glycopyrronium stearate is shown in 5 Fig. 27. Peak 1 shows potassium still present in the sample. Peak 2 is unknown, but was also present in the blank sample. As shown in Fig. 27, very little potassium remains in the sample, which indicates that the reaction was successful in producing the desired result, glycopyrronium stearate. The combination of the results shown in Fig. 26 and 27, taken together, also indicate a successful reaction.

Gas Chromatography for Fatty Acid Content

Another analytical method that was developed was a method for quantitation of weakly- or non-chromophoric starting materials and degradants (fatty acids and derived salts, dimethylhydroxypyrrolidinium degradants). The method used a GC (FID) method to determine stearic acid content.

Gas chromatography was used to determine stearic acid content in glycopyrronium stearate. A direct injection gas chromatography method was developed for Wt% assay for stearic acid. Glycopyrronium stearate is dissolved in THF and acidified with glacial acetic acid. The method is potentially applicable to other fatty acids (with modifications to the temperature gradient). Other components of the glycopyrronium fatty acid salts mixture were not detected (but therefore also not interfering). Sample analysis gave the results shown in Table 31.

Table 31

Sample	Description	Stearic Acid Content
EE-008-001-3a	Filterable Solids from GPBr / Stearic Acid / Potassium Stearate (recovered portion of excess free stearic acid)	98.3% wt.
EE-008-001-3b	Waxy Solids from GPBr / Stearic Acid / Potassium Stearate (mixture of GP stearate and excess free stearic acid)	51.45% wt.

WO 2016/204998

PCT/US2016/035967

It is somewhat difficult to separate the excess free fatty acid from the salt. In order to determine how much excess free fatty acid is still in the mixture, one can subtract the amount bound to glycopyrronium (which can be determined using stoichiometry calculations) to get the excess. The filterable solid in Table 31 represents what precipitates out, and is predominantly the free fatty acid (stearic acid in the table). This filterable solid, which is fairly pure (98.3% by weight of stearic acid) is discarded. The desired product (glycopyrronium stearate) is in the waxy solid. While some of the excess has been filtered out in the filterable solid, the stearic acid measured in the waxy solid is both the stearic acid that is part of glycopyrronium stearate and the excess free stearic acid.

While Figs. 9 and 10 and Table 31 only show gas chromatography results for stearic acid content, gas chromatography could also be used to quantitate other fatty acids, with minor modifications to the gas chromatography method. Some parameters that could be modified include, but are not limited to, the hold time, the injection temperature, the concentration for injection, and the time to ramp up the temperature. Alternatively, NMR results may be used to quantitate the amount of fatty acid. The amount of bound fatty acid can be determined using stoichiometry. The length of the particular fatty acids determines how much of the glycopyrronium fatty acid salt is fatty acid and how much is glycopyrronium. The gas chromatography operations are shown in Table 32.

Table 32

Parameter	Condition
Inlet	Split; Split Ratio 50:1
Inlet Temperature	250°C
Thermal Program	Initial 235°C (hold for 30 mins isothermal)
Detection	Flame Ionization
Detector Temperature	300°C
Carrier Gas	Helium

WO 2016/204998

PCT/US2016/035967

Makeup Gas	Helium, 30 mL/min
Air Flow	350mL/min
Hydrogen Flow	30 mL/min
Lit Offset	0.5
Mode	Constant flow
Flow	3.1 mL/min
Run Time	21 min

Fig. 28 shows a standard chromatogram and Fig. 29 shows a sample chromatogram. The approximate retention time for the peak ID for stearic acid was 20.1 minutes.

The complex mixtures that resulted from the chemical reactions to make the glycopyrronium fatty acid salts of interest were not amenable to straightforward analysis. For example, they included an oily mixture with free fatty acids. Most chromatographic methods use detection of a compound, but individual methods fell short of being able to capture the information about all of the materials (including starting materials, degradants, and final products). For example, UV can see the starting material and final product, but not the degradant QAA. The final product includes a chromophore, which is why it can be detected with UV, but the degradation product without the chromophore is not detectable with UV detection methods. Due to the chemical nature of the individual components, multiple analytical strategies were needed. Multiple methods were used, in tandem, to quantify the starting material, final product, and degradation products. Analysis may be optimized for each of the fatty acids used in the reactions.

Procedures were developed for glycopyrronium fatty acid salts and these procedures enabled the preparation of glycopyrronium fatty acid salts from a variety of fatty acids. Analysis of the fatty acid salts included various procedures. These procedures included HPLC with photodiode array detection, GC (FID) method for stearic acid,

WO 2016/204998

PCT/US2016/035967 bromide content by ion chromatography (anion mode), and potassium content by ion chromatography (cation mode). In addition to these methods, analysis by NMR (in d6DMSO) was found to be useful for the characterization of the isolated materials.

While the examples described herein predominantly discuss glycopyrronium 5 bromide, another quaternary ammonium anti-cholinergic muscarinic receptor antagonist, for example trospium chloride, could substitute for glycopyrronium bromide in the reaction to product a quaternary ammonium fatty acid salt (for example, a trospium fatty acid salt).

In one method of manufacturing such a salt, trospium chloride is reacted with a fatty acid salt in a biphasic reaction mixture, and at least 0.2 molar equivalent of excess free fatty acid is added to the reaction mixture. The method preferably mixes a fatty acid with water, 2-methyl-tetrahydrofuran (or another appropriate solvent), and a metal hydroxide selected from the group consisting of an alkali metal hydroxide and an alkaline earth metal hydroxide, until all solids are dissolved to form the fatty acid salt. Any of the steps and variations described with respect to the glycopyrronium fatty acid salts could alternatively be used to make trospium fatty acid salts.

All cited references are incorporated by references herein.

Accordingly, it is to be understood that the embodiments of the invention herein described are merely illustrative of the application of the principles of the invention.

Reference herein to details of the illustrated embodiments is not intended to limit the scope of the claims, which themselves recite those features regarded as essential to the invention.

WO 2016/204998

PCT/US2016/035967

Claims (38)

1. What is claimed is:

   1. A method of synthesizing at least one glycopyrronium fatty acid salt product comprising the step of mixing glycopyrronium bromide with a fatty acid salt in a biphasic reaction mixture, wherein at least a 0.2 molar equivalent of excess free

   5 fatty acid is added to the reaction mixture.
2. 2. The method of claim 1, wherein the mixing step comprises the substep of:

   i) mixing a fatty acid with water, 2-methyl-tetrahydrofuran, and a metal hydroxide selected from the group consisting of an alkali metal hydroxide and an alkaline earth metal hydroxide, until all solids are

   10 dissolved to form the fatty acid salt.
3. 3. The method of claim 2, wherein the mixing step further comprises the substeps of:

   ii) adding glycopyrronium bromide and mixing until solids dissolve;

   iii) removing a lower aqueous phase of the reaction mixture and retaining an upper organic phase;

   15 iv) washing the upper organic phase with water;

   v) repeating steps iii) and iv) at least once;

   vi) performing vacuum distillation on the upper organic phase;

   vii) adding new 2-methyl-tetrahydrofuran;

   viii) repeating step vi) until no distillate is observed;

   20 ix) adding a non-polar hydrocarbon solvent and filtering the resulting mixture to remove any suspended solids; and

   x) performing vacuum distillation of a filtrate from step ix) to obtain the product.

   WO 2016/204998

   PCT/US2016/035967
4. 4. The method of claim 3, wherein, in step v), substeps iii) and iv) are repeated at least two times.
5. 5. The method of claim 3, wherein the non-polar hydrocarbon solvent is selected from the group consisting of n-heptane, a mixture of heptane isomers, n-hexane, isooctane, and petroleum ether.
6. 6. The method of claim 1, wherein the product is selected from the group consisting of an oil, an oily solid, and a waxy solid.
7. 7. The method of claim 1, wherein a cation of the fatty acid salt is selected from the group consisting of Na, K, and Ca.
8. 8. The method of claim 1, wherein the fatty acid is selected from the group consisting of:

   arachidic acid, stearic acid, palmitic acid, oleic acid, erucic acid, linoleic acid, arachidonic acid, lauric acid, capric acid, linoleic acid, α-linolenic acid, γ-linolenic acid and myristic acid.
9. 9. The method of claim 1, wherein the fatty acid comprises at least eight carbon molecules.
10. 10. The method of claim 1, wherein an isolated glycopyrronium fatty acid salt mixture has an enrichment of the fatty acid relative to glycopyrronium compared to an input ratio.
11. 11. The method of claim 1, wherein at least a 0.6 molar equivalent of excess free fatty acid is added to the reaction mixture.
12. 12. The method of claim 1, wherein between approximately 0.6 molar equivalent of excess free fatty acid and 1.2 molar equivalent of excess free fatty acid is added to the reaction mixture.
13. 13. The method of claim 1, wherein at least a 1.1 molar equivalent of excess free fatty acid is added to the reaction mixture.

    WO 2016/204998

    PCT/US2016/035967
14. 14. The method of claim 1, wherein approximately 1.2 molar equivalent of excess free fatty acid is added to the reaction mixture.
15. 15. The method of claim 14, wherein a ratio of the fatty acid relative to glycopyrronium of an isolated glycopyrronium fatty acid salt mixture is between approximately 2.25:1 and 3.00:1.
16. 16. The method of claim 15, wherein a ratio of the fatty acid relative to glycopyrronium of an isolated glycopyrronium fatty acid salt mixture is between approximately 2.29:1 and 2.87:1.
17. 17. The method of claim 1, wherein the biphasic reaction mixture comprises water and 2methyl-tetrahydrofuran.
18. 18. The method of claim 1, further comprising the step of analyzing the glycopyrronium fatty acid salt product, comprising the substeps of:

    i) performing high pressure liquid chromatography to quantitate a glycopyrronium component and a hydrolysis degradant, Acid A, present in the product;

    ii) performing gas chromatography to quantitate a total fatty acid component in the product;

    iii) performing anion mode ion chromatography to quantitate a bromide component in the product;

    iv) performing cation mode ion chromatography to quantitate a potassium component in the product;

    v) performing stoichiometric calculations combining the results from substeps i), ii), iii), and iv) to determine an amount of excess free fatty acid and an amount of a quaternary amino alcohol degradant; and

    WO 2016/204998

    PCT/US2016/035967 vi) performing nuclear magnetic resonance to determine a molar ratio of glycopyrronium to total fatty acid in the product.
19. 19. A glycopyrronium fatty acid salt made by the method of claim 1, and having the chemical formula:

    <s .A.

    O' R wherein R is a fatty acid chain comprising at least eight carbon molecules.
20. 20. The glycopyrronium fatty acid salt of claim 19, wherein R is selected from the group consisting of:

    C11H23 (lauric acid);

    10 C17H35 (stearic acid);

    C17H33 (oleic acid);

    C15H31 (palmitic acid);

    C9H19 (capric acid);

    C19H31 (aradichonic acid);

    15 C19H39 (arachidic acid);

    C21H41 (erucic acid);

    C17H31 (linoleic acid);

    C17H29 (linolenic acid); and

    C13H27 (myristic acid).

    WO 2016/204998

    PCT/US2016/035967
21. 21. A method of synthesizing trospium fatty acid salts comprising the step of mixing trospium chloride with a fatty acid salt in a biphasic reaction mixture, wherein at least a 0.2 molar equivalent of excess free fatty acid is added to the reaction mixture.
22. 22. The method of claim 21, wherein the mixing step comprises the substep of:

    i) mixing a fatty acid with water, 2-methyl-tetrahydrofuran, and a metal hydroxide selected from the group consisting of an alkali metal hydroxide and an alkaline earth metal hydroxide, until all solids are dissolved to form the fatty acid salt.
23. 23. A compound with the chemical formula:

    ¹—CK

    N‘ o

    wherein R is a fatty acid chain comprising at least eight carbon molecules.
24. 24. The compound of claim 23, wherein R is selected from the group consisting of:

    C11H23 (lauric acid);

    C17H35 (stearic acid);

    C17H33 (oleic acid);

    C15H31 (palmitic acid);

    C9H19 (capric acid);

    C19H31 (aradichonic acid);

    C19H39 (arachidic acid);

    C21H41 (erucic acid);

    WO 2016/204998

    PCT/US2016/035967

    C17H31 (linoleic acid);

    C17H29 (linolenic acid); and

    C13H27 (myristic acid).
25. 25. A method of analyzing a glycopyrronium fatty acid salt product, comprising the steps

    5 of:

    a) performing high pressure liquid chromatography to quantitate a glycopyrronium component and a hydrolysis degradant, Acid A, present in the product;

    b) performing gas chromatography to quantitate a total fatty acid

    10 component in the product;

    c) performing anion mode ion chromatography to quantitate a bromide component in the product;

    d) performing cation mode ion chromatography to quantitate a potassium component in the product;

    15 e) performing stoichiometric calculations combining the results from steps

    a), b), c), and d) to determine an amount of excess free fatty acid and an amount of a quaternary amino alcohol degradant; and

    f) performing nuclear magnetic resonance to determine a molar ratio of glycopyrronium to total fatty acid in the product.

    WO 2016/204998

    PCT/US2016/035967

    1/38

    Fig. 1

    HPLC Calibration Curve for SM-2 by Bromide Area

    HPLC Area of Glycopyrronium (mAU*s)

    .......................................................................Fig. 2.......................................................................

    HPLC Calibration Curve for SM-2 by Glycopyrronium

    SUBSTITUTE SHEET (RULE 26)

    WO 2016/204998

    PCT/US2016/035967

    2/38

    Fig. 3

    Comments | Repeated analysis of water layer after standing in the separating funnel for 48 h for improved separation Concentrated at 25 °C under vacuum Repeated analysis of water layer after standing in the separating funnel for 48 h for improved separation PLC Results (AUC) | Glycopyrronium 88.96% | 93.99% 96.26% 96.93% 93.08% 91.64% 14.81 % | 21.79% | 28.39% | 83.33% | 75.77% | Acid A | 2.26% | 2.38% 2.39% 2.34% 5.18% 5.60% :%:% | 1.06% | | 2.40% | | 5.29% I | 5.08% | 9.64% X Bromide o' N CO 3.61% 1.34% 0.71% 1.73% 2.74% :::%¾ | 84.12% | 75.79% | 66.30% | 11.57% 14.58% | Sample Sample Details Main Me-THF layer Me-THF after first water wash Me-THF after second water wash Me-THF after third water wash Me-THF after third water wash After concentration %%% Main water layer First water wash Second water wash Third water wash Third water wash Q | XL-007-106-1 XL-007-106-2 XL-007-106-3 XL-007-1 06-4 XL-007-106-4 XL-007-106 :%:% | XL-007-1 06-1A | XL-007-106-2A | XL-007-106-3A | XL-007-106-4A XL-007-1 06-4A

    SUBSTITUTE SHEET (RULE 26)

    WO 2016/204998

    PCT/US2016/035967

    3/38

    Fig. 4

    Comments | Concentrated at 25 °C under vacuum HPLC Results (AUC) Glycopyrronium 93.08% 95.05% 94.83% 88.31% 1 \° iff 6^s ;;; 00 | O ss 20.32% 25.88% 70.23% Acid A 0.92% | 1.43% 2.75% 3.13% none none none 5.03 | Bromide 5.99% | 3.50% 2.41% 8.55% ;;; C 1 5) 1 LO SS 00 79.67% | 74.11% | 24.73% | | Sample Sample Details Me-THF after first water wash Me-THF after second water wash Me-THF after third water wash After concentration Main water layer First water wash Second water wash Third water wash g | XL-007-109-1 | XL-007-109-2 XL-007-109-3 XL-007-109 V t-601--ΖΌ0-ΊΧ | XL-007-109-2A | | XL-007-109-3A | | XL-007-109-4A |

    SUBSTITUTE SHEET (RULE 26)

    WO 2016/204998

    PCT/US2016/035967

    4/38

    Fig. 5

    Reaction 4 | 2.4 g 2.8 g (4.0 eq.) f-Amyl alcohol I 9 o CO 14 hours I Seven hours: 51.8 area% of SM-1 30.4 area% of Acid A 17.8 area% of methyl ester Fourteen hours 24.5 area% of SM-1 30.3 area% of Acid A 45.2 area% of methyl ester | V/N N/A N/A Methylation in f-amyl alcohol at -100°C with extended reaction time failed to afford the desired quaternary product. By-product Acid a and its methyl ester was formed durinq the reaction Reaction 3 2.5 g 2.3 g (3.0 eq.) DMAc co 9 o CO 7 hours Seven hours: 3.6 area% of SM-1 12.4 area% of Acid A 84 area% of methyl ester 2.2 g after aqueous wash -95 area% of methyl ester Product was confirmed as methyl ester of Acid A Methylation in DMAc at 130C failed to provide the desired quaternary product, but SM-1 was converted to acid A and its methyl ester. Reaction 2 1.0 g 1.18 g(4.0 eq.) f-Amyl alcohol cT 9 o St 2 hours N/A 1.2 g after concentration -85 area% of SM-1 -2.3 area% of “Glycopyrronium” -11 area% of Acid A The methyl peaks corresponding to the quaternary ammonium salt were very small Methylation in f-amyl alcohol at -140°C over two hours failed to afford complete methylation, but hydrolysis of SM-1 to Acid A occurred. Reaction 1 1.0 g 5.0 g(16.8 eq.) No extra solvent 9 o CO 6 hours N/A 1.1 g after concentration No “Glycopyrronium” peak was observed and some un-reacted SM-1 was present No methyl peaks matching the quaternary ammonium salt Methylation in excess dimethyl carbonate failed to provide the desired methylated product. φ w co 0 co ο ΆΟ σίτ-O CO ' OJ 0^0) CC co 2 I Solvent Oil bath temperature I Reaction Time IPC Results I Product HPLC Results LO ω Ξ (Λ Φ CC CC Σ z T Conclusions

    0) 6 ’w w Q. Q. LO LO

    .z _z 4—1 4—1 s £ Φ Φ L_ 3 3 w ω ω w Φ Φ k_ Ω. Q. X X CC CO E E co co o o 4—¹ 4—¹ Ό 0 Φ Φ W ω CO co Φ Φ 0 0 z z

    Φ Φ i— =5 =5 w ω cn ω Φ Φ o_ a co co z z cd ω

    4—I -I—I z z

    CD CD .Z -Z Η H

    The internal pressure increased to a max pressure with 60 psig.

    The internal pressure increased to a max pressure with 25 psig.

    Acid A was confirmed with the sample independently prepared by hydrolysis of SM-1.

    04 CO LO

    SUBSTITUTE SHEET (RULE 26)

    WO 2016/204998

    PCT/US2016/035967

    5/38

    Fig. 6

    SUBSTITUTE SHEET (RULE 26)

    WO 2016/204998

    PCT/US2016/035967

    6/38

    AREA PERCENT REPORT

    SORTED BY: SIGNAL

    MULTIPLIER: 1.0000

    DILUTION: 1.0000

    USE MULTIPLIER & DILUTION FACTOR WITH ISTDs

    SIGNAL 1: DAD1 A, SIG=210,8 REF-360,100

    PEAK RET TIME TYPE WIDTH AREA HEIGHT AREA # i (MIN) ! (MIN) (mAU*s) _: {mALD % 1 ¹ 15.161 MM 0.1272 2089.22754 273.65390 96.0360 2 15.619 MF 0.0567 3.57500®-] l.O5047e-l 0.0164 3 15.705 MF 0.0746 7.69562®-! 1.71847®-! 0.0354 4 15.827 FM 0.1645 4.08049 4.13368®-! 0.1876 5 17.779 MM 0.0781 5.85760®-] 1.24941®-! 0.0269 6 18.982 MM 0.1080 8.85946®-! 1.36721®-! 0.0407 7 20.325 MM 0.0839 3.42095e-] 6.79694®-2 0.0157 8 20.495 MM 0.0937 10.65661 1.09477 0.4899 9 21.099 MM 0.0771 2.62635®-! 5.67742e-2 0.0121 10 21.887 MM 0.1132 1.08475 1.59675®-! 0.0499 11 27.447 MM 0.0815 2.17781 4.45613®-! 0.1001 12 28.632 MM 0.2774 22.36731 1.34365 1.0202 13 29.313 MM 0.0963 38.52335 6.66453 1.7700 14 31.643 MM 0.1509 1.46514 1.61774e-l 0.0673 15 32.174 MF 0.1355 7.88104®-! 9.69468e-2 0.0362 16 32.306 FM 0.1488 9.89743®-! 1.10865®-! 0.0455 17 33.912 MM 0.1541 8.97962®-! 9.71092®-! 0.0413 rOIALS: 2175.46240 285.70550

    FIS. 7

    SUBSTITUTE SHEET (RULE 26)

    WO 2016/204998

    PCT/US2016/035967

    7/38

    AREA PERCENT REPORT

    SORTED BY: SIGNAL

    MULTIPLIER: 1.0000

    DILUTION: 1.0000

    USE MULTIPLIER & DILUTION FACTOR WITH ISTDs

    SIGNAL 1: BAD] A, SIG-210,8 REF-360,100

    FEAR RET TIME TYPE WIDTH AREA HEIGHT AREA # (MIN) i (MIN) (mAU*s) (mAU) % 1 ¹ 15.146 MM 0.1268 2072.02051 ¹ 272.32730 ¹ 87.6482 2 15.673 MF 0.1331 7.89762®-! 9.88960®-2 0.0334 3 15.848 FM 0.1611 2.03526 2J0608e~l 0.0861 4 16.223 MF 0.1604 7,31523®-] 7.23953®-2 0.0309 5 16.330 FM 0.1058 3.49233©-! 5.50007e-2 0.0148 6 16.647 MF 0,0652 2.50569©-] 6.403978-2 0.0106 7 16.703 FM 0.0944 3.25546©-! 5.750428-2 0.0138 8 17.773 MM 0.0759 6.06932©-! 1.333138-1 0.0257 9 18.989 MM 0.0904 5.99195©-] 1.104968-1 0.0253 10 20.319 MF 0.0776 3.81165e-1 8.18591 g-2 0.0161 11 20.400 FM 0.0028 224.44206 45.17399 9.4941 12 21.098 MM 0.0825 2.98388©-] 6.0274U-2 0.0126 13 21.885 MM 0.0960 8.75967©-] 1.521568-1 0.0371 14 27.429 MM 0.0755 2.54322 5.61148g-l 0.1076 15 29.295 MM 0.0976 50.23679 8.57705 2.1251 16 31.633 MM 0.1830 3.00260 2.734868-1 0.1270 17 32.167 MF 0.1980 2.88648 2.42939e-l 0.1221 18 32.293 FM 0.1333 1.64481 2.056008-1 0.0696 TOTALS: 2364.02080 328.45806

    FIG. 8

    SUBSTITUTE SHEET (RULE 26)

    WO 2016/204998

    PCT/US2016/035967

    8/38

    SUBSTITUTE SHEET (RULE 26)

    WO 2016/204998

    PCT/US2016/035967

    9/38

    SUBSTITUTE SHEET (RULE 26)

    WO 2016/204998

    PCT/US2016/035967

    10/38

    SUBSTITUTE SHEET (RULE 26)

    WO 2016/204998

    PCT/US2016/035967

    11/38

    SUBSTITUTE SHEET (RULE 26)

    WO 2016/204998

    PCT/US2016/035967

    12/38

    SUBSTITUTE SHEET (RULE 26)

    WO 2016/204998

    PCT/US2016/035967

    13/38

    SUBSTITUTE SHEET (RULE 26)

    WO 2016/204998

    PCT/US2016/035967

    14/38

    SUBSTITUTE SHEET (RULE 26)

    WO 2016/204998

    PCT/US2016/035967

    15/38

    SUBSTITUTE SHEET (RULE 26)

    WO 2016/204998

    PCT/US2016/035967

    16/38

    SUBSTITUTE SHEET (RULE 26)

    WO 2016/204998

    PCT/US2016/035967

    17/38

    SUBSTITUTE SHEET (RULE 26)

    WO 2016/204998

    PCT/US2016/035967

    18/38

    SUBSTITUTE SHEET (RULE 26)

    WO 2016/204998

    PCT/US2016/035967

    19/38

    SUBSTITUTE SHEET (RULE 26)

    WO 2016/204998

    PCT/US2016/035967

    20/38

    SUBSTITUTE SHEET (RULE 26)

    WO 2016/204998

    PCT/US2016/035967

    21/38

    AREA PERCENT TOT

    SORTED BY: SIGNAL

    CALIB. DATA MODIFIED: 12/05/2014 0:40:34 AM

    MULTIPLIER: 1.0000

    DILUTION: 1.0000

    USE MULTIPLIER & DILUTION FACTOR WITH ISTDs

    PEAK # RET TIME (MIN) SIGNAL 1: BAD! A, SIG=210,8 REF=360,l00 TYPE I I WIDTH [MINI AREA (mAU*s) AREA % NAME 1 ^! 13.050 ¹ BB¹ 0.1404 1469.35474 ^S 90.8652 ¹ GP 2 13.788 BV 0.1333 1.38567 0.0857 1 3 15.008 BY 0.1086 3.38840e-l 0.0210 i 4 15.339 BV 0.0173 3.19566b-! 0.0190 ? 5 15.830 8V 0.0797 5.53090e~1 0.0342 ? 6 17.165 BV 0.0936 6.71909e-l 0.0416 ? 7 17.682 BB 0.0807 92.99374 5.750B ACID A 8 20.177 BV 0.0606 6.81594e-1 0.0421 2 9 24.393 BB 0.0950 27.90519 1.7306 ? 10 24.815 BB 0.0880 2.03513 0.1259 i 11 26.398 BB 0.0798 20.75115 1.2B33 ?

    TOTALS: 1617.07061

    FIG 158

    SUBSTITUTE SHEET (RULE 26)

    WO 2016/204998

    PCT/US2016/035967

    22/38

    AREA PERCENT REFORT

    SORTED BY: SIGNAL

    CALI8, DATA MODIFIED: 12/05/2014 0:40:34 AM

    MULTIPLIER: 1.0000

    DILUTION: 1.0000

    USE MULTIPLIER & DILUTION FACTOR WITH ISTDs

    PEAK # RETIME (MINI SIGNAL!: DA01 A, SIG=210,0 REF=360,100 TYPE ί I WIDTH (MINI AREA (mAITs) AREA % 1 NAME 1 ^! 13.050 BB 0.1426 1480.29688 90.0661 ¹ GP 2 13.766 RV 0.1456 1.53003 0.0939 1 3 15.003 BY 0.0859 4.999398-1 0.0307 i 4 15.313 BB 0.0469 1653978-1 0.0102 7 5 15.828 BV 0.0777 5.954208-1 0.0365 ? 6 17.161 BV 0.1103 9.000808-1 0.0602 ? 7 17.684 BB 0.0788 94.28443 5.7875 ACID A 0 20.179 BV 0.0707 4.704458-1 0.0294 7 9 24.395 BB 0.0905 26.70000 16309 7 10 24,018 BB 0.0B57 1,94646 0.1195 7 ]] 26.399 BB 0.0807 21.61063 13270 ?

    TOTALS: 162109649 fig. 150

    SUBSTITUTE SHEET (RULE 26)

    WO 2016/204998

    PCT/US2016/035967

    23/38

    SUBSTITUTE SHEET (RULE 26)

    WO 2016/204998

    PCT/US2016/035967

    24/38

    AREA PERCENT TOT

    SORTED 8Y:

    CALIB. DATA MODIFIED: MULTIPLIER:

    DILUTION:

    SIGNAL

    12/05/20! 4 8:40:34 AM

    1.0000

    1.0000

    USE MULTIPLIER & DILUTION FACTOR WITH ISTDs

    SIGNAL 1: DAD1 A, SIG=210,8 REF=360,l 00 PEAK # RET TIME _; (MIN) TYPE I 1 WIDTH WIN) AREA (mAiTs) AREA % .........................................................1 NAME :............................................... 1 13.058 bb ¹ 0.1414 1372.55225 ¹ 91.2525 ^{1 !} GP 2 13.824 BBA 0.1648 1.08466 0.0721 ? 3 15.855 BV 0.0893 6.35602s-] 0.0423 ? 4 17.133 BV 0.0946 8.06691-3-1 0.0536 2 5 17.686 BV 0.0808 61.94052 4.1180 ACID A 6 18.309 BB 0.0713 4.62513e-1 0.0307 2 1 18.587 BV 0.0856 6.09179©-! 0.0405 ? 0 20,178 BV 0.0632 3.29590e-l 0.0219 2 9 24.477 BB 0.0903 3.14137 0.2089 2 10 24,816 BB 0.0834 3.01253 0.2003 2 11 26,401 BB 0.0807 30.61319 2.0353 2 12 28.566 BB 0.1318 4.86617 0.3235 ? 13 29.584 BB 0.1172 24.07088 1.6003 7

    TOTALS: 1504.12514

    FIS. 161

    SUBSTITUTE SHEET (RULE 26)

    WO 2016/204998

    PCT/US2016/035967

    25/38

    AREA PERCENT REPORT

    SORTED BY: SIGNAL

    CALIB. DATA MODIFIED: 12/05/2014 0:40:34 AM

    MULTIPLIER: 1.0000

    DILUTION: 1.0000

    USE MULTIFLIER & DILUTION FACTOR WITH ISTDs

    SIGNAL 1: DAD1 A_# SIG=2I0,8 REF=360,l 00 PEAK RETIME TYPE WIDTH (MIN) _i AREA (mAU*s) AREA % NAME # (MIN) 1 ¹ 13.057 ¹ BB ¹ 0.1408 ³ 1363.68787 ^! 91.3890 GP 2 13.796 BBA 0.1060 5.88608®-] 0.0394 ? 3 15.022 BB 0.0907 6.96373e-I 0.0467 ? 4 17.172 BV 0.1278 1.11755 0.0749 7 5 17.690 BB 0.0811 59.03640 4.0100 ACID A 6 10.313 BV 0.0636 4.38142®-! 0.0294 ? 1 10.576 BV 0.0930 9.2010k·] 0.0617 7 8 20.103 BV 0.0545 4.27755®-] 0.02B7 7 9 24.477 BB 0.0937 2.97541 0.1994 7 10 24.818 BB 0.0054 3.15553 0.2115 7 11 26.399 BB 0.0804 29.99472 2.0101 7 12 28.509 BB 0.1291 3.75733 0.2518 ? 13 29.572 BB 0.1162 24.58340 1.6475 7

    TOTALS: 1492.17926

    FIG. 16D

    SUBSTITUTE SHEET (RULE 26)

    WO 2016/204998

    PCT/US2016/035967
26. 26/38

    SUBSTITUTE SHEET (RULE 26)

    WO 2016/204998

    PCT/US2016/035967
27. 27/38

    SORTED BY: SIGNAL

    CALI B. DATA MODI RED: 12/05/2014 8:40:34 AM

    MULTIPLIER: 1.0000

    DILUTION: 1.0000

    USE MULTIPLIER & DILUTION FACTOR WITH ISTDs

    SIGNAL 1: DAD1 A, 5IG-210,8 REF=36Q,100

    PEAK RETTIME TYPE WIDTH (MIN) AREA (mAU*s) AREA % ! NAME # . (MIN) _! I ¹ 13.070 ^! BB ¹ 0.1352 ¹ 1163.46033 ¹ 39.7889 ¹ GP 2 15.865 BB 0.0939 4.13036®-! 0.0141 3 17.136 BB 0.O432 2.14958®-! 7.35! e-3 2 4 17.692 BB O.0808 70.54691 2.4126 ACID A 5 18.309 BBA 0.0662 2.93541 e-1 0,0100 9 6 18.989 BV 0.0796 6.31166®-! 0.0216 2 7 20.184 BV 0.0616 3.60141®-! 0.0123 2 8 20.386 W 0.1058 1.23050 0.0421 ? 9 21.261 BV 0.1090 2.03984 0.0698 9 10 22.699 BV 0.1379 8.42927 0.2883 2 II 23.359 BV 0.1343 2.18818 0.0748 2 12 24.01 B BB 0.1154 2.24744 0.0769 2 13 26.401 BV 0.1177 23,02224 0.7873 2 14 26.653 VB 0.1116 2.62104 0.0896 9 15 27.853 BV 0.0890 1614.01575 55.1974 2 16 28.158 W 0.1018 22.51083 0.7698 2 17 28.344 VB 0.1657 6,38885 0.2185 9 IB 32.591 BB 0.1287 3.46753 0.1186 2

    TOTALS: 1492.17926

    FIG. 17B

    SUBSTITUTE SHEET (RULE 26)

    WO 2016/204998

    PCT/US2016/035967
28. 28/38

    AREA PERCENT REPORT

    SORTED BY:

    CALIB. DATA MODIFIED:

    MULTIPLIER:

    DILUTION:

    SIGNAL

    12/05/2014 8:40:34 AM

    1.0000

    1.0000

    USE MULTIPLIER & DILUTION FACTOR WITH ISTDs

    SIGNAL 1: DAD1 A, SIG-210,8 REF-36G,100

    PEAK REFTIME TYPE WIDTH (MIN) AREA imAITs) AREA % I NAME # ί BIN) _; 1 ¹ 13.051 ^! BB ^! 0.1399 ¹ 1402.12244 39.5620 ^! GP 2 14.848 BV 0.0628 3.66435e-1 0.0103 ? 3 15.814 BV 0.0718 4.62627e-1 0.0131 ? 4 17.142 BV 0.1082 1.15670 0.0326 ? 5 17.686 BB 0.0787 85.66539 2.4171 ACID A 6 20.175 BV 0.0541 3,89841 @4 0.0110 1 1 20.382 W 0.1030 1.34319 0.0379 ? B 21.247 BV 0.1020 1.83370 0.0517 ? 9 22.692 BV 0.1103 6,94507 0.1960 10 22.846 W 0.0717 1.10283 0.0311 7 11 23.355 BV 0.0004 1.22437 0.0345 ? 12 24.476 BB 0.0930 1,55670 0.0439 ? 13 24.813 BBA 0.0790 1.31747 0.0372 ? 14 26.396 BV 0.0746 15,82830 0.4466 ? 15 26.466 W 0.0711 12.21716 0.3447 ? 16 26,651 VB 0.1450 4,21305 0.1189 ? 17 27.848 BV 0.0893 1964,56592 55.4318 ? 18 28.155 W 0.1058 28.56983 0.8061 ? 19 28.335 VB 0.1547 8,38243 0.2365 ? 20 32.606 BB 0.1398 4.84739 0.1368 ?

    TOTALS: 3544.11171

    HG. 170

    SUBSTITUTE SHEET (RULE 26)

    WO 2016/204998

    PCT/US2016/035967
29. 29/38

    SUBSTITUTE SHEET (RULE 26)

    WO 2016/204998

    PCT/US2016/035967
30. 30/38

    SUBSTITUTE SHEET (RULE 26)

    WO 2016/204998

    PCT/US2016/035967
31. 31/38

    SUBSTITUTE SHEET (RULE 26)

    WO 2016/204998

    PCT/US2016/035967
32. 32/38

    SUBSTITUTE SHEET (RULE 26)

    WO 2016/204998

    PCT/US2016/035967
33. 33/38

    GP Cone, vs GP Peak Area

    o o o O o o o O o o o o sj- m f\J t—1

    eajv>|eatj dD

    SUBSTITUTE SHEET (RULE 26)

    WO 2016/204998

    PCT/US2016/035967
34. 34/38

    SUBSTITUTE SHEET (RULE 26)

    WO 2016/204998

    PCT/US2016/035967
35. 35/38

    SUBSTITUTE SHEET (RULE 26)

    WO 2016/204998

    PCT/US2016/035967
36. 36/38

    SUBSTITUTE SHEET (RULE 26)

    WO 2016/204998

    PCT/US2016/035967
37. 37/38

    SUBSTITUTE SHEET (RULE 26)

    WO 2016/204998

    PCT/US2016/035967
38. 38/38

    SUBSTITUTE SHEET (RULE 26)